VYSIS ALK BREAK APART FISH PROBE KIT, VYSIS PARAFFIN PRETREATMENT IV & POST HYBRIDIZATION WASH BUFFER KIT PROBECHEK ALK

{0} SUMMARY OF SAFETY AND EFFECTIVENESS DATA (SSED) I. GENERAL INFORMATION Device Generic Name: DNA FISH Probe Assay Device Trade Name: Vysis ALK Break Apart FISH Probe Kit, Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit, ProbeChek ALK Negative Control Slides, and ProbeChek ALK Positive Control Slides Applicant's Name and Address: Abbott Molecular, Inc. 1300 E. Touhy Avenue Des Plaines, IL 60018 Date(s) of Panel Recommendation: None Premarket Approval Application (PMA) Number: P110012 Date of FDA Notice of Approval: August 26, 2011 Expedited: Granted expedited review status on April 21, 2011 because it was believed through literature support that the Vysis Anaplastic Lymphoma Kinase (ALK) Break Apart FISH Probe Kit would aid in the selection of previously treated non-small cell lung cancer patients for treatment with the Pfizer drug Xalkori® (crizotinib), which may result in a life-prolonging survival benefit and is therefore in the best interest of those patients. Because no other legally marketed device is available, the FDA decided to grant expedited review to the Vysis ALK Break Apart FISH Probe Kit. II. INDICATIONS FOR USE The Vysis ALK Break Apart FISH Probe Kit is a qualitative test to detect rearrangements involving the ALK gene via fluorescence in situ hybridization (FISH) in formalin-fixed, paraffin-embedded (FFPE) non-small cell lung cancer (NSCLC) tissue specimens to aid in identifying patients eligible for treatment with Xalkori® (crizotinib). This is for prescription use only. The Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit is used to prepare paraffin-embedded lung cancer tissue sections fixed on positively charged slides for use in fluorescence in situ hybridization (FISH) with Vysis DNA FISH probes. The ProbeChek ALK Negative Control Slides are intended for use as an assay control for appropriate hybridization conditions during routine use of the Vysis ALK Break Apart FISH Probe Kit (List No. 06N38-020). The ProbeChek ALK Negative Control Slides should be assayed in conjunction with the user's specimen slides according the package insert for the Vysis ALK Break Apart FISH Probe Kit (List No. 06N38-020). PMA P110012: FDA Summary of Safety and Effectiveness Data {1} The ProbeChek ALK Positive Control Slides are intended for use as an assay control for appropriate hybridization conditions during routine use of the Vysis ALK Break Apart FISH Probe Kit (List No. 06N38-020). The ProbeChek ALK Positive Control Slides should be assayed in conjunction with the user’s specimen slides according the package insert for the Vysis ALK Break Apart FISH Probe Kit (List No. 06N38-020). ## III. CONTRAINDICATIONS None ## IV. WARNINGS AND PRECAUTIONS The warnings and precautions can be found in the Vysis ALK Break Apart FISH Probe Kit, Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit, ProbeChek ALK Negative Control Slides, and ProbeChek ALK Positive Control Slides labeling. ## V. DEVICE DESCRIPTION The Vysis ALK Break Apart FISH Probe Kit uses FFPE tissue sections which have been mounted to glass slides as the specimen. The tissue sections are deparaffinized and then treated with components of the Vysis Paraffin Pretreatment IV &amp; Post-Hybridization Wash Buffer Kit. The DNA contained within the nuclei of the FFPE tissue sections is denatured to the single-stranded form and subsequently allowed to hybridize with the locus-specific indicator Vysis ALK Break Apart FISH Probes. Following hybridization, the unbound probe is removed by a series of washes using Vysis Wash Buffers I and II and the nuclei are counterstained with DAPI Counterstain (4,6 diamidino-2-phenylindole), a DNA-specific stain that fluoresces blue. Hybridization of the Vysis ALK Break Apart FISH Probes is viewed using a fluorescence microscope equipped with appropriate excitation and emission filters, allowing visualization of the orange and green fluorescent signals. The Vysis ALK Break Apart FISH Probe Kit assay requires four kits: - Vysis ALK Break Apart FISH Probe Kit - Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit - ProbeChek ALK Negative Control Slides - ProbeChek ALK Positive Control Slides ## The Vysis ALK Break Apart FISH Probe Kit: The Vysis ALK Break Apart FISH Probe Kit consists of one 200 μL vial of Vysis LSI ALK Dual Color Break Apart FISH Probe and a 300 μL vial of DAPI I Counterstain. It contains sufficient reagents to process 20 assays. Each assay uses 10 μL of Vysis ALK Break Apart FISH Probe Kit applied to a hybridization target area of 22 mm x 22 mm. The probe requires storage at -30°C to -10°C upon receipt. The VYSIS ALK BREAK APART FISH PROBE KIT, as shown in Figure 1, is a mixture that consists of two DNA probes directly labeled with fluorophores in hybridization buffer containing dextran sulfate, formamide, and SSC with blocking DNA: - Vysis LSI 3’-ALK SpectrumOrange - Vysis LSI 5’-ALK SpectrumGreen PMA P110012: FDA Summary of Safety and Effectiveness Data {2}  Figure 1. Probe map and ideogram for the Vysis ALK Break Apart FISH Probe Kit probes.  The hybridization targets of these probes are on opposite sides flanking the breakpoint region of the ALK gene. The 3'-ALK probe that hybridizes telomerically of the breakpoint is approximately 300 kb and is labeled with the SpectrumOrange (SO) fluorophore. The 5'-ALK probe that hybridizes centromerically of the breakpoint is approximately 442 kb and is labeled with the SpectrumGreen (SGn) fluorophore. **Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit:** The Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit consists of five 50 mL bottles of pretreatment solution (1N Sodium thiocyanate); five bottles Protease Buffer IV (0.1N hydrochloric acid); five bottles Protease IV (Pepsin 2500-4000 U/mg); one 250 mL bottle of Wash Buffer I (0.3% NP-40/0.7XX SSC pH 7); and one 250 mL bottle of Wash Buffer II (0.1% NP-40/2X SSC, pH 7.0). The Protease IV reagent requires storage at -20°C to ±10°C upon receipt and one vial the Protease is added to the contents of one bottle of Protease IV Buffer on the day of use. All other reagents are provided ready-to-use. **ProbeChek ALK Positive and Negative Control Slides:** A negative control (ProbeChek ALK Negative Control Slides) and positive control (ProbeChek ALK Positive Control Slides) are provided separately with five slides per kit. Both controls consist of FFPE cultured cell line specimens which have been sectioned and applied onto glass microscope slides. The cell line used for the negative control shows no chromosomal rearrangement at the ALK breakpoint and provides ALK signal enumeration consistent with a negative classification. The cell line mixture used for the positive control provides ALK signal enumeration consistent with a positive classification (20 - 62%). Control slides must be run concurrently with patient slides to monitor assay performance and to assess accuracy of signal enumeration. Control slides are introduced at the initiation of the slide deparaffinization procedure and evaluated to determine assay validity. Control slides are run on each day of FISH testing and with each new kit lot. **Specimen Preparation:** The Vysis ALK Break Apart FISH Probe Kit has been optimized only for identifying and quantifying rearrangements of the ALK gene from FFPE human NSCLC tissue specimens. The assay should be performed only on 10% neutral buffered formalin FFPE human lung tumor tissue. Other types of specimens or fixatives should not be used. PMA P110012: FDA Summary of Safety and Effectiveness Data {3} # Signal Enumeration: For each nucleus, the number of fused (adjacent) signals, single orange signals, and single green signals are recorded. An individual cell is counted only once regardless of the number of rearrangements and/or deletions that it contains. Users are instructed to not score nuclei with no signals or with signals of only one color (without a fused and/or broken apart signal). Only those nuclei with one or more FISH signals of each color are scored. A nucleus that contains signals that are weak or overly diffuse should not be enumerated. Cells are considered negative (non-rearranged) when: - Orange and green signals are adjacent or fused. Orange and green signals that are less than two signal diameters apart are considered as a single fused signal. - There is a single green signal without a corresponding orange signal. Cells are considered positive (rearranged) when: - At least one set of orange and green signals are two or more signal diameters apart. - There is a single orange signal without a corresponding green signal in addition to fused and/or broken apart signals. Table 1. Classification of Cells as Positive or Negative | No. of Adjacent or Fused Signals | No. of Single Orange Signals | No. of Single Green Signals | Cell Classification | | --- | --- | --- | --- | | ≥1 | 0 | 0 | Negative | | ≥1 | 0 | ≥1 | Negative | | ≥0 | ≥1 | ≥1 | Positive | | ≥1 | ≥1 | 0 | Positive | The number of cells classified as positive or negative for ALK rearrangements or deletions are calculated according to the table above. - A sample is considered negative if &lt;5 cells out of 50 (&lt;5/50 or &lt;10%) are positive. - A sample is considered positive if &gt;25 cells out of 50 (&gt;25/50 or &gt;50%) are positive. - A sample is considered equivocal* if 5–25 cells (10–50%) are positive. If the sample is equivocal, a second reader should evaluate the slide. - The first and second cell count readings are added together and a percent is calculated out of 100 cells (average percent of positive cells). - If the average percent positive cells is &lt;15% (&lt;15/100), the sample is considered negative. - If the average percent positive cells are ≥15% (≥15/100), the sample is considered positive. *Specimens whose results fall into the equivocal zone (10 – 50% positive), based on the enumeration by the first reader, should be enumerated by a second reader to confirm results. # VI. ALTERNATIVE PRACTICES AND PROCEDURES There are currently no alternative approved methods to the Vysis ALK Break Apart FISH Probe Kit assay for detecting ALK rearrangements in NSCLC. PMA P110012: FDA Summary of Safety and Effectiveness Data {4} # VII. MARKETING HISTORY The Vysis ALK Break Apart FISH Probe Kit, ProbeChek ALK Negative Control Slides, and ProbeChek ALK Positive Control Slides have not been marketed in the United States or any foreign country. The Vysis Paraffin Pretreatment IV &amp; Post Hybridization Wash Buffer Kit has not been marketed in the United States but is marketed in Europe. # VIII. POTENTIAL ADVERSE EFFECTS OF THE DEVICE ON HEALTH Failure of the device to perform as expected or failure to correctly interpret test results may lead to incorrect ALK test results and subsequently improper patient management decisions in NSCLC treatment. For the specific adverse events that occurred in the clinical studies, please see the Adverse Event section the approval documents for NDA 202570. # IX. SUMMARY OF PRECLINICAL STUDIES Analytical performance was assessed to evaluate the safety and effectiveness of the Vysis ALK Break Apart FISH Probe Kit assay on NSCLC tissue. The studies included probe concentration optimization, hybridization quality evaluation, establishment of the normal cutoff, probe localization, FISH success rate, analytical sensitivity, analytical specificity, microbial contamination, robustness, kit shelf-life stability, reproducibility, and repeatability testing. The robustness studies included: slide thickness, slide stability (pre-hybridization and post-hybridization), water bath section flotation temperature and time, slide baking, deparaffinization, pretreatment conditions, enzyme digestion, probe and target co-denaturation, hybridization time and temperature, post-hybridization wash, and DAPI staining. ## A. Laboratory Studies ### Evaluation of hybridization quality: In all studies, the quality of the hybridization of each probe is evaluated for each hybridization target. Each hybridization target area within the assay is rated on a scale of 1 to 5 according to four attributes: probe signal intensity, nuclear and chromosomal specificity, target background, and cross-hybridization. An overall rating is then assigned to each slide based upon the lowest score of the attributes evaluated. For each slide, an overall rating of greater than or equal to three (3) indicates a disposition of “Pass”, otherwise the disposition is “Fail.” ### Probe Concentration Optimization: The Vysis ALK Break Apart FISH Probe Kit is composed of two individual probes. In order for the probe to be enumerated properly, the signal intensities of the two probes need to be balanced so the intensity (brightness) of the two probes is optimal for visualization. Two NSCLC FFPE tissue specimens were processed according to the package insert which included pretreatment, denaturation, hybridization and post-hybridization washes. Four readers were utilized to evaluate the results according to the qualitative/quantitative assessments described above with signal intensity being the primary variable. PMA P110012: FDA Summary of Safety and Effectiveness Data page 5 9 {5} The concentration of the 5'ALK SGn probe was evaluated at three different starting concentrations. The 5'ALK SGn probe was fixed at one concentration after no significant difference was noted between the concentrations tested. The 3'ALK SO Probe concentration was added and tested at using various different concentrations. Quality scores with respect to intensity ≥ 3 were considered acceptable. One concentration of the 3'ALK SO probe was selected as it yielded the optimal balance of probe intensities with the intensity quality score of both the SGn and SO probes of ≥3. Once the two probes were optimized, the probe was tested using slides from 16 individual ALK-positive and negative NSCLC FFPE blocks and two lots of ProbeChek ALK Control slides (in duplicate). The 5'ALK SGn probe had 93% (56/60) success rate and 3'ALK SO Probe had 97% (58/60) success rate. ## Normal Cutoff The normal cutoff value was defined as the maximum amount of scoreable interphase nuclei with a specific abnormal signal pattern at which a specimen is considered negative for that signal pattern. The normal cutoff value is expressed in terms of a percentage, or the actual number of nuclear FISH patterns positive for rearrangement per the standard number of nuclei tested. The normal cutoff was established as 15% using NSCLC FFPE tissue specimens. ## Probe Localization on Metaphase Chromosomes To verify that the probes hybridized only to the intended locus, the location of hybridization of the Vysis ALK Break Apart FISH Probe Kit was evaluated on eight metaphase spreads prepared from cultured lymphocyte slide preparations in conjunction with the inverted DAPI chromosome banding technique. Both probes which make up the final probe mixture were shown to hybridize to the intended locus (2p23) on all eight metaphase spreads and to no other locations. ## Analytical Sensitivity and Specificity Analytical sensitivity was defined as the percentage of chromosome targets with the expected normal signal pattern. Analytical specificity was defined as the percentage of signals that hybridize to the correct locus and no other location. The sensitivity and specificity of the LSI 3'-ALK SpectrumOrange and LSI 5'-ALK SpectrumGreen FISH probes were evaluated using metaphase chromosomes prepared from 6 slide lots created from peripheral blood cultures of karyotypically normal specimens from 5 individual donors. The analytical sensitivity was calculated to be 100.0% (240/240) (95% CI 98.5-100.0%) for each probe. The sensitivity calculation, the signals for LSI 3'-ALK SO and LSI 5'-ALK SGn FISH probes, were enumerated for each metaphase spread (normal = 2 signals). In total, 240 signals were expected for each probe (2 signals per cell x 20 metaphase spreads per lot x 6 slide lots). For the specificity calculation, the number of metaphase spreads with the expected signal pattern was enumerated. In total, 120 metaphase spreads were evaluated (20 metaphase spreads x 6 slide lots). The analytical specificity was calculated to be 100.0% (120/120) (95% CI 97.0-100.0%) for each probe. PMA P110012: FDA Summary of Safety and Effectiveness Data {6} # FISH Success Rate on FFPE NSCLC Specimens The FISH success rate on FFPE lung tumor specimens was evaluated using FFPE tumor tissue specimens. The assay success rates when using a dual bandpass green/orange filter v2 and the single bandpass green was 99.2% (119/120) and the lower-bound of the 95% confidence interval was 96.1%. The assay success rates for the single bandpass orange were 98.3% (118/120) and the lower-bound of the 95% confidence interval was 94.8%. # Microbial Contamination The Vysis ALK Break Apart FISH Probe Kit met the requirements for a microbiologically uncontrolled product per “Guideline for the Manufacture of In Vitro Diagnostic Products”, 1/10/1994, as none of the reagents would sustain growth of the selected microorganisms and in fact killed the applied inoculum of microorganisms as referenced by the lack of growth upon subculture. Additionally upon testing the reagents in the normal QC procedure, all the reagents performed satisfactorily even after three days of incubation with the selected organisms at 35-37°C. # Robustness Studies The objective of the guard banding studies was to establish the robustness of the Vysis ALK Break Apart FISH Probe Kit for detection of ALK gene rearrangements by FISH in NSCLC FFPE tissues. These studies included evaluation of specimen preparation (section thickness and water bath temperature), specimen pretreatment (slide baking, deparaffinization, pepsin digestion time), probe and target co-denaturation (time and temperature), wash conditions (wash solutions, time, temperature), DAPI staining, and slide stability (Pre-Hybridization and Post-Hybridization) using both NSCLC FFPE tissue samples and cell lines (ProbeChek ALK Control Slides). For all studies, a qualitative assessment of hybridization and/or a quantitative assessment of fluorescent signals were used to determine the acceptable guard banding condition. A brief description of each characterization/robustness study is described below. For all, the interpretation of the FISH results included both qualitative assessment of hybridization and quantitative assessment of fluorescent signals. All slides were processed through all assay steps and evaluated. A slide was considered of acceptable quality (passing) if the overall qualitative score was ≥ 3. A slide was classified as “Fail” if either one of two readers gave it an overall quality score of &lt; 3. 1. Section Thickness – Section thickness of the FFPE tissue was evaluated within the range of variability for slide preparation as thickness of the FFPE tissue sections may affect signal enumeration results for the Vysis ALK Break Apart FISH Probe Kit due to nuclear truncation or the degree of cell overlap. The three NSCLC specimens consisted of two ALK-positives and one ALK-negative and each specimen was sectioned into 4 μm, 5 μm and 6 μm thicknesses. Six sections were cut for thickness and the slides were blinded, processed, and enumerated by two readers, resulting in a total of 108 PMA P110012: FDA Summary of Safety and Effectiveness Data {7} enumerations (3 specimens, 3 section thicknesses, 6 slides, 2 readers). The percentage of ALK-positive cells was determined for each slide. The statistical significance of the analyses was defined at the 5% level and variance assessed by ANOVA. From the results, there is little statistical impact on the positive cell enumeration for the Vysis ALK Break Apart FISH Probe Kit assay between 4 – 6 μm and therefore the recommended thickness is 5 ± 1 μm. 2. Slide Stability – Pre-Hybridization - Archived tissue specimens stored for approximately 2 years were evaluated with respect to assay performance to determine if up to 2 years of storage time is acceptable for this assay. Twenty (20) sections from 4 different NSCLC FFPE tissue specimen blocks were sectioned at 5 μm and tested between 23.5 – 24.5 months. The slides were enumerated by two independent readers. There were a total of 40 individual readings (20 per reader). Out of 40 readings, 36 (90%) indicated an overall quality passing rate. 3. Sample Preparation - Water bath Section Flotation Temperature – NSCLC FFPE tissue specimens and control slides were processed for up to 30 minutes and evaluated using eight flotation water bath conditions at four temperatures ranging between 37°C to 50°C and evaluated for flotation times of &lt;1 min and 30 min at each temperature and evaluated by two readers. There was no difference seen in sections which remained in the flotation water bath for 30 minutes at 37-50°C. 4. Slide Baking - Prior to deparaffinization, NSCLC FFPE tissue sections are baked to assure adherence of FFPE sections to the slides. A guardbanding study was conducted using the ThermoBrite to characterize the slide baking step of the assay on NSCLC FFPE tissue specimens. NSCLC FFPE tissue specimens were each processed and evaluated using four slide baking conditions: 56°C for 2 hours, 60°C for 2 hours, 60°C for 24 hours, and 68°C for 24 hours. All of the slides evaluated resulted in scores with acceptable hybridization quality scores. 5. Deparaffinization (Hemo-De) Exposure Time – The purpose of this study was to evaluate a range of exposure times of the NSCLC FFPE tissue slides to Hemo-De (xylene substitute) as a deparaffinization pretreatment process. NSCLC FFPE tissue specimens and cell lines (ALK Negative and Positive Controls) were processed and evaluated using three deparaffinization conditions times. All slides were read by two independent readers. All of the slides evaluated resulted in scores with acceptable hybridization quality scores. 6. Pretreatment Conditions – Pretreatment is utilized to reverse the effects of cross-linking induced by fixation, to make the tissue permeable, to remove cytoplasm, and to digest proteins in order to make the genomic target DNA in the nuclei assessable for hybridization, as well as to reduce background PMA P110012: FDA Summary of Safety and Effectiveness Data page 8 12 {8} autofluorescence of cells and tissue. The pretreatment procedures used in FISH involve the application of high heat, chaotropic agents such as sodium thiocyanate, and protein digestion. Following deparaffinization, the cross-links formed during tissue fixation are broken by incubating sections at high temperature with sodium thiocyanate followed by protein digestion by a protease (pepsin). The effect of the temperature and time of incubation in Pretreatment Solution (1N sodium thiocyanate, NaSCN) on tissue morphology and hybridization signal quality were evaluated using the Vysis ALK Break Apart Probe Kit. NSCLC FFPE tissue specimens and cell lines were each processed and evaluated using three pretreatment solution incubation conditions: - 9 minutes at 77°C - 12 minutes at 80°C - 15 minutes at 83°C All slides were evaluated by two independent readers. All of the slides evaluated resulted in scores with acceptable hybridization quality scores. 7. Enzyme Digestion (Pepsin Concentration and Time) – Enzyme concentration is important for the performance of the assay, since inadequate enzymatic digestion of tissue could lead to weak signal and background autofluorescence, while over-digestion could damage nuclear morphology, and decrease signal intensity. NSCLC FFPE tissue specimens and cell lines were each processed and digested for 20 minutes at various concentrations to determine the optimal digestion times. All slides were evaluated by two independent readers and a pepsin concentration range which demonstrated acceptable hybridization quality was determined. Digestion time was evaluated using a specific pepsin concentration (which fell within the acceptable range) at 16, 18, 20, 22, and 24 minutes at 37°C. All slides were evaluated by two independent readers. In this experiment, greater than 85% of evaluations for each time point were necessary to “Pass.” In addition, the percentage of tissue loss (as assessed by the percent of degraded nuclear morphology) was evaluated in this experiment. If the tissue had greater than 25% degraded nuclear morphology it was considered failing. In this experiment, all conditions tested resulted in less than 10% tissue loss per specimen slide. Three replicates (two from one ALK-negative and one from an ALK-positive) received a minimum quality score of two at the 20 minute time point. This resulted in an overall quality score indicating a failure at this time point; however more than 85% of the evaluations still passed the quality evaluations and resulted in less than 10% degradation in nuclear morphology, therefore it was considered to still pass. 8. Probe and Target Co-Denaturation (Denaturation Study) – Co-denaturation of probe and sample DNA, and hybridization of the probe to its genomic target are carried out simultaneously on the ThermoBrite hybridization platform PMA P110012: FDA Summary of Safety and Effectiveness Data page 9 13 {9} upon probe application to the target area of the slide. Denaturation time and temperature for probe and target DNA is critical for subsequent hybridization. NSCLC FFPE tissue specimens were each processed and evaluated using seven denaturation conditions: - 3 min. at 71°C - 3 min. at 73°C - 3 min. at 75°C - 5 min. at 71°C - 5 min. at 73°C - 5 min. at 80°C - 2 min. at 73°C All slides were evaluated by two independent readers. A trend towards decreased signal intensity and specificity was observed at 80°C and 75°C × 5 min, and a trend towards decreased signal intensity and higher background was observed at 71°C × 3 min. Therefore, assay limits were found to be minimum of 71°C × 5 min and maximum of 75°C × 3 min. This study was repeated using five denaturation temperatures: - 5 min. at 71°C - 3 min. at 73°C - 3 min. at 75°C - 2 min. at 73°C - 5 min. at 73°C All slides were evaluated by two independent readers. All of the slides evaluated resulted in scores with acceptable hybridization quality scores. The denaturation condition was demonstrated to result in acceptable hybridization quality for all 5 conditions. 9. Hybridization Time and Temperature – Hybridization time and temperature for the FFPE lung tissue digestion were tested with Vysis ALK Break Apart FISH Probe Kit on NSCLC FFPE tissue specimens and cell lines. Each specimen was hybridized at three different temperatures (34°C, 37°C and 40°C) and three time points at each temperature (14 hours, 19 hours and 24 hours) to determine acceptable ranges of conditions for probe hybridization on the Thermobrite Platform. An additional condition (14 hours at 36°C) was tested with the same specimen set to further define the acceptable range of hybridization conditions. Denaturation conditions were held constant at 73°C × 3 min. The slides were evaluated by two different readers, and increments of 0.5 were used to increase the discriminatory power of the quality evaluation. Hybridization for 14 hours at 34°C (low temperature and time) resulted in average overall quality ratings of &lt;3 and was therefore excluded from the acceptable range of the assay. Table 2. Hybridization Time and Temperature Results. | Temp. (°C) | Time (hrs) | Average Quality Rating (min, max) | No. Passing Results (≥3) | No. Failing Results (<3) | | --- | --- | --- | --- | --- | | 34° | 14 | 2.83 (2.0, 4.0) | 8 | 4 | | | 19 | 3.00 (2.0, 4.0) | 10 | 2 | | | 24 | 3.50 (2.0, 4.0) | 11 | 1 | | 36° | 14 | 3.13 (2.0, 4.0) | 11 | 1 | PMA P110012: FDA Summary of Safety and Effectiveness Data {10} | Temp. (°C) | Time (hrs) | Average Quality Rating (min, max) | No. Passing Results (≥3) | No. Failing Results (<3) | | --- | --- | --- | --- | --- | | 37° | 14 | 3.17 (2.5, 4.0) | 11 | 1 | | | 19 | 3.17 (2.5, 4.0) | 11 | 1 | | | 24 | 3.08 (2.5, 4.0) | 10 | 2 | | 40° | 14 | 3.08 (3.0, 3.5) | 12 | 0 | | | 19 | 3.04 (2.0, 4.0) | 11 | 1 | | | 24 | 3.08 (3.0, 3.5) | 12 | 0 | An additional hybridization condition, 14 hours at 36°C, was tested with the same specimen set to establish the lower temperature/time bound of the acceptable range. The hybridization condition with the low limit of 14 hours at 36°C and high limit of 24 hours at 40°C were determined to be acceptable. 10. Post-Hybridization Wash – Post-hybridization wash is employed after the completion of hybridization and consists of two washes, a room-temperature wash (to remove coverslips) and a high-temperature wash. In some assay protocols an additional room temperature wash is used. A stringency of the high temperature wash is critical for the removal of the un-specifically bound probe and unbound probe. The factors affecting the stringency of the high-temperature wash are salt concentration, temperature and time. These factors were evaluated in characterization and guard band experiments. The studies were performed in two stages with the hybridization temperature kept constant at 37°C for the entire experiment. The first stage used two denaturation conditions and three wash conditions, with four NSCLC FFPE 5 µm tissue specimens (three ALK-negatives and one ALK-positive) and two lots of the ProbeChek ALK Negative Control. Three readers each evaluated six slides yielding 18 results per condition. The slides after pretreatment, hybridization and post-hybridization processing using an internal quality procedures with increments of 0.5, were used to increase discriminatory power of assay quality evaluation. Wash 1 condition remained constant for all conditions at 2X SSC/0.1% NP40. An additional room temperature wash following the high temperature 2X SSC/0.1% NP40 wash at 73°C for 2 minutes was determined not to be required for the Vysis ALK Break Apart FISH Probe Kit. As post-hybridization wash and denaturation are critical steps affecting probe intensity, specificity and background, the second stage of the experiment evaluated the best condition from the stage one. The second stage utilized four NSCLC FFPE tissue specimens (ALK-negative), two ALK-positive cultured cell lines, and one lot of the ProbeChek ALK Positive Control which was evaluated by three independent readers to confirm the condition selected. Additionally, for the slides washed at high temperature in 2X SSC/0.1% NP40, the 0.7X SSC/0.3% NP40 was used as the PMA P110012: FDA Summary of Safety and Effectiveness Data {11} room temperature wash. The use of 0.7X SSC/0.3% NP40 as the room temperature wash condition was incorporated into the final protocol. Increments of 0.5 were used to increase discriminatory power of assay quality evaluation. As shown in Table 3, Condition 5 (73°C x 3 min denaturation followed by 74°C, 2 min, 2X SSC/0.1% NP40 wash and 0.7X SSC/0.3% NP40 room temperature wash) was superior to the starting condition in overall quality scores and the percentage individual readings that passed hybridization quality assessment (overall quality rating of ≥3). Table 3. Denaturation and Wash Condition Optimization Stage 2. | Condition | Denaturation | Wash 1 Condition (ambient) | Wash 2 Condition (high temperature) | Sample type | Avg. Quality Rating (min, max) | % Passing (≥3) | | --- | --- | --- | --- | --- | --- | --- | | 1 | 71°C x 5 min | 2X SSC/ 0.1% NP40 | 73°C x 2 min in 0.7X SSC/0.3% NP40 | FFPE cell line | 2.92 (2.00, 4.00) | 66.67 | | 5 | 73°C x 3 min | 0.7X SSC/ 0.3% NP40 | 74°C x 2 min in 2X SSC/0.1% NP40 | FFPE cell line | 3.25 (2.00, 4.00) | 83.33 | | 1 | 71°C x 5 min | 2X SSC/ 0.1% NP40 | 73°C x 2 min in 0.7X SSC/0.3% NP40 | NSCLC Tissue | 2.64 (2.00, 3.50) | 58.33 | | 5 | 73°C x 3 min | 0.7X SSC/ 0.3% NP40 | 74°C x 2 min in 2X SSC/0.1% NP40 | NSCLC Tissue | 3.11 (2.00, 4.00) | 86.11 | A guard band study was performed to assess the acceptable range surrounding the high temperature wash conditions. A range of wash temperatures (72 – 76°C) and times (1 – 6 min.) with NSCLC FFPE specimens and cell lines were used in the evaluation by three readers. Increments of 0.5 were used to increase discriminatory power of assay quality evaluation. From the data, signal quality decreased at for the 6 minute time point at 74°C and all time points for the 76°C temperature. More than 90% of the evaluations were successful for post-hybridization wash conditions between 72 – 74°C for 1 – 4 minutes. A final confirmatory experiment at the selected conditions deemed to be optimal, 2 min at 74°C and the temperature range of 73°C – 75°C using NSCLC FFPE tissue specimens and cell lines were used in the evaluation by twp readers. The results confirmed the recommended wash range of 2 minutes at 74±1°C as all slides tested passed hybridization quality assessment by both readers, and all of the overall hybridization quality scores were acceptable. 11. DAPI Staining Time Course –FFPE NSCLC specimens were stained with DAPI/anti-fade solution which was added to the slides (1 at a time), coverslipped, and the time was started. The slides were photographed with the camera attached to a microscope at 1 minute intervals until 10 minutes, then 5 minute intervals to 25 minutes to assess saturation of DAPI staining in a pictorial record. By visual assessment without the aid of photography, 2 minutes post-DAPI staining appeared to be sufficient to visualize tissue morphology and nuclear boundaries. No change in DAPI intensity was PMA P110012: FDA Summary of Safety and Effectiveness Data {12} detected after 7 minutes, even when compared to a slide at 25 – 27 minutes post-staining or greater. Morphology could be assessed with ease at 5 – 6 minutes post-DAPI staining. Staining reached saturation at 15 – 20 minutes by visualization of the photographic images. No reduction in intensity after 25 minutes was noted. 12. Slide Stability: Post-Hybridization – Signal stability after hybridization was evaluated since fluorescence signals will fade over time. The FISH hybridization success rate on FFPE lung tumor specimens was evaluated using thirty different FFPE tumor tissue specimens. Each FFPE specimen block was sectioned into four sections, each 5 µm in thickness, and mounted on a glass slide. The specimens were processed and stored at -20°C for 5, 6, and 7 days. The FISH success rate on FFPE was evaluated for the dual and single bandpass filter sets. Results are shown below in Tables 4 and 5. Based on the data, the slide stability was demonstrated for a maximum of seven days post-hybridization when stored at -20°C. Table 4. Post-hybridization storage stability – overall. | Filter Set | # Hybridizations | # Hybridizations with Passing Overall Quality Rating (≥3) | Assay Success Rate (%) | Lower Bound 95% Confidence Limit | | --- | --- | --- | --- | --- | | Dual bandpass green/orange v2 | 120 | 119 | 99.2 | 96.1 | | Single bandpass green | 120 | 119 | 99.2 | 96.1 | | Single bandpass orange | 120 | 118 | 98.3 | 94.8 | Table 5. Post-hybridization storage stability – by day. | Filter Set | Percentage of Passing Results (%) | | | | --- | --- | --- | --- | | | Day 5 n=72 slides | Day 6 n=28 slides | Day 7 n=20 slides | | Dual bandpass green/orange v2 | 98.6 (71/72) | 100 (28/28) | 100 (20/20) | | Single bandpass green | 98.6 (71/72) | 100 (28/28) | 100 (20/20) | | Single bandpass orange | 97.2 (70/72) | 100 (28/28) | 100 (20/20) | ## Conclusions The following results were observed to be acceptable: - Section thickness for 4 µm, 5 µm, and 6µm. - Slide stability (FFPE tissue section) pre-hybridization for up to 24 months. - Flotation water bath temperature at 37°C to 50°C during specimen preparation. - Slide baking 56°C to 68°C for 2 hours to 24 hours. - Deparaffinization in Hemo-De for three (3) 5-minute exposures. - Pretreatment at 75°C to 85°C for 9 min. to 15 min. - Pepsin digestion time for 16 min. to 24 min. - Denaturation for 5 min. at 71°C (lower limit) and 3 min. at 75°C (upper limit). - Hybridization 36°C to 40°C for 14 to 24 hours. PMA P110012: FDA Summary of Safety and Effectiveness Data {13} - Post-Hybridization Wash Solution 2 conditions at 72°C to 76°C for 1 – 4 min. - DAPI staining for a minimum of 3 min. - Slide stability post-hybridization for up to 7 days at -20°C (±10°C). Note: It is recommended to adhere to the time and temperatures indicated in the staining procedure provided in the package insert. ## Reproducibility 1. Control Slide Reproducibility – Control slide reproducibility was evaluated using three lots of both the ProbeChek ALK Negative Control Slides and ProbeChek ALK Positive Control Slides. Each lot was run on 5 non-consecutive days over a 23-day time period and evaluated by three readers for a total of 90 data points (3 lots x 5 runs x 3 readers = 45 evaluations per control slide type). For each specimen, the signal patterns of 50 nuclei were evaluated by counting the number of fused signals, single orange signals and single green signals present for each target by each reader. The control slides are described in Tables 6 and 7. Table 6. ProbeChek ALK Positive Control Slides Descriptive Statistics | Control Slide Lot | N | Reader 1 | | | | Reader 2 | | | | Reader 3 | | | | Total | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Mean | SD | Min | Max | Mean | SD | Min | Max | Mean | SD | Min | Max | n | Mean | SD | Min | Max | | 1 | 5 | 41.6 | 9.32 | 28 | 54 | 40.4 | 5.90 | 32 | 48 | 50.4 | 21.61 | 34 | 76 | 15 | 44.1 | 13.76 | 28 | 76 | | 2 | 5 | 36.4 | 5.90 | 28 | 44 | 42.4 | 2.61 | 38 | 44 | 46.8 | 23.35 | 24 | 74 | 15 | 41.9 | 13.68 | 24 | 74 | | 3 | 5 | 46.0 | 9.59 | 36 | 56 | 45.6 | 4.98 | 38 | 50 | 50.8 | 21.94 | 28 | 74 | 15 | 47.5 | 13.30 | 28 | 74 | | Total | 15 | 41.3 | 8.80 | 28 | 56 | 42.8 | 4.89 | 32 | 50 | 49.3 | 20.74 | 24 | 76 | 45 | 44.5 | 13.47 | 24 | 76 | Mean= mean percentage of cells with ALK rearrangements SD = standard deviation between percentage of cells with ALK rearrangements Min = minimum percentage of cells with ALK rearrangements Max = maximum percentage of cells with ALK rearrangements Table 7. ProbeChek ALK Negative Control Slides Descriptive Statistics. | Control Slide Lot | n | Reader 1 | | | | Reader 2 | | | | Reader 3 | | | | Total | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Mean | SD | Min | Max | Mean | SD | Min | Max | Mean | SD | Min | Max | n | Mean | SD | Min | Max | | 1 | 5 | 4.0 | 2.83 | 2 | 8 | 1.2 | 1.10 | 0 | 2 | 1.2 | 1.79 | 0 | 4 | 15 | 2.1 | 2.33 | 0 | 8 | | 2 | 5 | 3.6 | 3.29 | 0 | 8 | 1.2 | 1.10 | 0 | 2 | 1.2 | 1.10 | 0 | 2 | 15 | 2.0 | 2.27 | 0 | 8 | | 3 | 5 | 2.0 | 1.41 | 0 | 4 | 1.2 | 1.79 | 0 | 4 | 1.6 | 1.67 | 0 | 4 | 15 | 1.6 | 1.55 | 0 | 4 | | Total | 15 | 3.2 | 2.60 | 0 | 8 | 1.2 | 1.26 | 0 | 4 | 1.3 | 1.45 | 0 | 4 | 45 | 1.9 | 2.04 | 0 | 8 | Mean= mean percentage of cells with ALK rearrangements SD = standard deviation between percentage of cells with ALK rearrangements Min = minimum percentage of cells with ALK rearrangements Max = maximum percentage of cells with ALK rearrangements There was no statistical difference in FISH classification between 3 readers by the Fisher-Freeman-Halton test at the significance level of 0.05 for either the ProbeChek ALK Positive or Negative Control slides as shown in Tables 8 and 9. Therefore, it was demonstrated that ProbeChek ALK Negative Control Slides and ProbeChek ALK Positive Control Slides could be reproducibly classified. All slides in this study were found to be within specifications. PMA P110012: FDA Summary of Safety and Effectiveness Data page 14 18 {14} Table 8. Reproducibility of ProbeChek ALK Positive Control Slides-By Reader | Readers | # Observations with the % ALK Rearrangement | | Total | | --- | --- | --- | --- | | | Within Specification* (≥20%) | Outside Specification (<20%) | | | 1 | 15 | 0 | 15 | | 2 | 15 | 0 | 15 | | 3 | 15 | 0 | 15 | | Fisher-Freeman-Halton p-value = 1.00 | | | | * Specification refers to the lower limit for the range of ALK rearrangements observed in the positive control cell line. Table 9. Reproducibility of ProbeChek ALK Negative Control Slides-By Reader | Readers | # of Observations with the % ALK Rearrangement | | Total | | --- | --- | --- | --- | | | Within Specification* (≤8%) | Outside Specification (>8%) | | | 1 | 15 | 0 | 15 | | 2 | 15 | 0 | 15 | | 3 | 15 | 0 | 15 | | Fisher-Freeman-Halton p-value = 1.00 | | | | * Specification refers to the upper limit for the range of ALK rearrangements observed in the negative control cell line. Some variation in the percentage of cells with the ALK rearrangement for within-readers (day-to-day), between-readers, and between-lots is evident. However, the variability did not affect the slide classification as demonstrated by the primary analysis of reader reproducibility and a secondary analysis of reproducibility by lot. All positive control slides were correctly classified. 2. Tissue Reproducibility – Tissue reproducibility was designed to evaluate the reader component of reproducibility by evaluation of between-reader and within-reader using FFPE lung tumor sections. This study was conducted using six serial sections (5 µm) prepared from twenty NSCLC FFPE specimen blocks. The panel included three ALK-positive specimens with &gt;50% of the cells with ALK rearrangement, three specimens falling within the range of 10% to 50% cells with the ALK rearrangement and fourteen ALK-negative specimens with &lt;10% cells with the ALK rearrangement. Three readers participated in the three study arms: between reader (Arm 1A), within reader (Arm 2), and between slide reproducibility (Arm 1B). The Vysis ALK Break Apart FISH Probe Kit was shown to be reproducible based upon the between-reader and between-slide analyses resulting in a Fisher-Freeman-Halton p-value of 1.00. a. Between Reader Reproducibility (Arm 1A) – Three readers evaluated the same set of slides prepared from 20 patient specimens resulting in a total of 60 evaluations. Each evaluation involved two enumerations of the same randomized, blinded slide by each reader for the purpose of determining the final classification (e.g., positive or negative) via averaging of the two enumerations. The final classification (negative or positive) was used for primary data analysis using kappa statistic. PMA P110012: FDA Summary of Safety and Effectiveness Data page 15 {15} This statistic measures the degree to which interpretation variability arises from differences among panel members relative to differences among readers interpreting the same panel member. For each panel member, the proportion of all possible pairings on which readers agree is calculated. An overall mean agreement of all panel members is determined and a correction for chance agreement is made to then determine a final kappa coefficient (K) and standard error (SE). Based on literature, an interpretation for the strength of agreement was assigned for various ranges of kappa. To show statistical significance, a test for the null hypothesis that the kappa coefficient is not different from zero (i.e., no better than chance), the generalized kappa statistic is compared with the standard normal distribution by creating a Z-Score (K/SE) and compared to 1.96 (alpha level of 0.05) for significance. An analysis of between reader reproducibility, based on the final FISH classification, with and without the equivocal calls is shown in Table 10 from a secondary analysis. This analysis used the initial classification, the first reading out of two for each slide by each reader, resulting in a positive, negative or equivocal classification. Table 10. Between-reader reproducibility – Classification | Not including “equivocal” results | Kappa Statistic: 1.00 (0.66, 1.00), SE = 0.171 Strength: Almost Perfect, z-score: 5.84 | | --- | --- | | Including “equivocal” results | Kappa Statistic: 0.72 (0.41, 1.00), SE = 0.162 Strength: Substantial, z-score: 4.46 | The Fisher-Freeman-Halton test was also used for reproducibility analyses of the data. The null hypothesis for this test showed no association between two variables as displayed in an r x c contingency table, where r is the number of rows and c is the number of columns. The alternative hypothesis (if the null is rejected with a significant p-value) shows that there is some statistically significant association, i.e., some effect, of the row variable on the column variable. The test was performed at a 0.05 level of significance. Table 11 shows the reproducibility between readers with the inclusion of the equivocal zone results. When the equivocal results are resolved into positive or negative results the Fisher-Freeman-Halton p-value was 1.00. Table 11. Between-Reader Reproducibility – Fisher-Freeman-Halton Analysis | | FISH Classification Number of Panel Members | | | | | --- | --- | --- | --- | --- | | Reader | Negative | Equivocal | Positive | Total | | 1 | 13 | 3 | 4 | 20 | | 2 | 14 | 2 | 4 | 20 | | 3 | 14 | 4 | 2 | 20 | Fisher-Freeman-Halton p-value: 0.85 b. Between-Slide Reproducibility (Arm 1B) – For this study, one reader was selected at random out of the three readers in the comparison between PMA P110012: FDA Summary of Safety and Effectiveness Data {16} readers (Arm 1A) and evaluated two additional serial sections from each of the patient specimens evaluated in Arm 1A (20 patient specimens, 2 slides each). The two additional evaluations were combined with the evaluation from the same reader for those patient specimens in Arm 1A. The resulting three evaluations from the same reader for each patient specimen yielded a total of sixty evaluations. Each evaluation involved two enumerations of the same randomized, blinded slide by the same reader for the purpose of determining the final classification (e.g. positive or negative) via averaging of the two enumerations. Table 12 shows the reproducibility between slides with the inclusion of the equivocal zone results. The final classification (negative or positive) was used for primary data analysis. A secondary analysis was conducted using the initial classification, the first reading out of two for each slide by each reader, resulting in a positive, negative or equivocal classification. When the equivocal results are resolved into positive or negative results the Fisher-Freeman-Halton p-value was 1.00. Table 12. Between-Slide Reproducibility – Fisher-Freeman-Halton Analysis. | | FISH Classification – Number of Panel Members | | | | | --- | --- | --- | --- | --- | | Slide | Negative | Equivocal | Positive | Total | | 1 | 14 | 2 | 4 | 20 | | 2 | 15 | 4 | 1 | 20 | | 3 | 14 | 4 | 2 | 20 | Fisher-Freeman-Halton p-value: 0.60 c. Within-Reader Reproducibility (Arm 2) – This study assessed reproducibility of the evaluation and classification of the same slide hybridized with the Vysis ALK Break Apart FISH Probe Kit over successive reading by the same reader. Three slides from twenty patient specimens were hybridized with the Vysis ALK Break Apart FISH Probe Kit and each of the three readers evaluated one slide exclusively three times in a random, blinded setup. To determine the final classification of samples (e.g., positive or negative), the following pairs of enumerated results were averaged together: 1) readings A and B, 2) readings B and C, and 3) readings A and C. This resulted in three evaluation results per slide for a total 180 evaluations. The final classification (negative or positive) was used for primary data analysis. A secondary analysis was conducted using the initial classification, the first reading out of two for each slide by each reader, resulting in a positive, negative or equivocal classification. For each arm (Arm 1A, Arm 1B, Arm 2), two-by-two contingency tables were constructed with the Final FISH classification (Positive, Negative) as the row, the agreement status (Agree, Disagree) as the columns, and the cells representing the number of results for each specimen which are found to agree or disagree. Positive Percent Agreement (PPA) and PMA P110012: FDA Summary of Safety and Effectiveness Data {17} Negative Percent Agreement (NPA) were calculated with corresponding 95% confidence intervals. With the exception of the PPA and OA for Arm 1B (Between-slides/within one reader), all other comparisons had an NPA of 100% (14/14) (CI95% 78.47, 100) and PPA was 100% (6/6) (CI95% 60.97, 100) and an overall percent agreement (OA) of 100% (20/20) (CI95% 83.89, 100). The PPA and OA for Arm 1B was 83.33% (5/6) (CI95% 43.65, 96.99) and the OA was 95% (19/20) (CI95% 76.39, 99.11). ## External Reproducibility Reproducibility of the Vysis ALK Break Apart FISH Probe Kit was evaluated at three external laboratories by testing a coded, randomized 12-member specimen panel (6 unique specimens, 2 slides each) that consisted of four unique ALK-positive with varying levels of positivity (Panel Member 1, 2, 3, and 6) and two unique ALK-negative NSCLC FFPE tissue specimens (Panel Member 4 and 5). Three lots of the Vysis ALK Break Apart FISH Probe Kit reagents were used in the evaluation. A run consisted of one replicate each of a ProbeChek ALK Negative Control slide, a ProbeChek ALK Positive Control slide and each panel member. Each of the three clinical sites tested the Reproducibility Panel using two of the three clinical lots. Lot 1 was tested at sites 1 and 3; Lot 2 at sites 1 and 2; and Lot 3 at sites 2 and 3. Each of the two technologists at each of the three testing sites enumerated the 6 study specimens and control slides once a day, for 5 non-consecutive days, per reagent lot over a period of 20 days. Each site evaluated 120 specimen slides for a total of 360. This resulted in 240 enumerations at each site for a minimum of 720 enumerations. Each site evaluated 40 controls slides (20 positive and 20 negative slides) for a total of 120. This resulted in 80 enumerations at each site for a minimum of 240 enumerations. For each panel member and control slides, the signal patterns of 50 nuclei were enumerated by two readers. An analysis of the data in Table 13 resulted in an overall kappa coefficient of 0.92 (95% CI 0.85 – 0.98) and standard error of 0.034. The Z-Score of 27.08, which is greater than 1.96, showed the kappa coefficient is significantly different from zero at a 0.05 level of significance. The kappa coefficient demonstrated the reproducibility for each site, ranging from 0.83 – 0.96, and for each lot, ranging from 0.86 – 0.96. The results are found in Tables 14 and 15, respectively. Table 13. Final FISH Classification – Kappa Analysis for Overall Reproducibility. | Panel Member | Number of Slides Across Sites/Lot/Runs/Readers | | | | --- | --- | --- | --- | | | Negative | Positive | Total | | 1 | 1 | 59 | 60 | | 2 | 0 | 60 | 60 | | 3 | 2 | 58 | 60 | | 4 | 60 | 0 | 60 | | 5 | 60 | 0 | 60 | | 6 | 4 | 56 | 60 | PMA P110012: FDA Summary of Safety and Effectiveness Data page 18 {18} Table 14. Final FISH Classification - Kappa Analysis, Reproducibility by Site | Site | Panel Member | Final FISH Classification No. of Slides Across Sites/Lot/Runs/Readers | | Kappa Statistics | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Negative | Positive | Kappa | CI95% | Strength | Standard Error | Z-score | | 1 | 1 | 0 | 20 | 0.96 | 0.83, 1.00 | Almost Perfect | 0.068 | 14.21 | | | 2 | 0 | 20 | | | | | | | | 3 | 0 | 20 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 1 | 19 | | | | | | | 2 | 1 | 0 | 20 | 0.96 | 0.83, 1.00 | Almost Perfect | 0.068 | 14.21 | | | 2 | 0 | 20 | | | | | | | | 3 | 0 | 20 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 1 | 19 | | | | | | | 3 | 1 | 1 | 19 | 0.83 | 0.72, 0.94 | Almost Perfect | 0.056 | 14.90 | | | 2 | 0 | 20 | | | | | | | | 3 | 2 | 18 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 2 | 18 | | | | | | Table 15. Final FISH Classification - Kappa Analysis, Reproducibility by Lot | Lot | Panel Member | Final FISH Classification No. of Slides Across Sites/Lot/Runs/Readers | | Kappa Statistics | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | Negative | Positive | Kappa | CI95% | Strength | Standard Error | Z-score | | 1 | 1 | 0 | 20 | 0.86 | 0.75, 0.98 | Almost Perfect | 0.059 | 14.75 | | | 2 | 0 | 20 | | | | | | | | 3 | 0 | 20 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 2 | 18 | | | | | | | 2 | 1 | 0 | 20 | 0.96 | 0.83, 1.00 | Almost Perfect | 0.068 | 14.21 | | | 2 | 0 | 20 | | | | | | | | 3 | 0 | 20 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 1 | 19 | | | | | | | 3 | 1 | 1 | 19 | 0.93 | 0.80, 1.00 | Almost Perfect | 0.065 | 14.34 | | | 2 | 0 | 20 | | | | | | | | 3 | 0 | 20 | | | | | | | | 4 | 20 | 0 | | | | | | | | 5 | 20 | 0 | | | | | | | | 6 | 1 | 19 | | | | | | PMA P110012: FDA Summary of Safety and Effectiveness Data {19} There was no significant association between the sites for the overall analysis (p-value of 0.8354) and by panel member (p-value ranged from 0.3220 to 1.00) by Fisher-Freeman-Halton analysis on the final FISH classification. For the analyses on the individual FISH classifications, there was a significant association between the sites for the overall analysis (p-value of &lt;0.0001) and for the positive panel members 1, 2, 3, and 6 (all p-values &lt;0.0001) and no significant association for the negative panel members 4 and 5 (both p-values of 1.00). The results for final FISH classification and individual FISH classification overall and individual FISH classification for each panel member are shown in Table 16. There was a statistically significant difference between sites in terms of reproducibility based on the raw results before any resolution of equivocal results. When equivocal results are resolved, the p-value of the Fisher-Freeman-Halton statistic changes to 0.8354, which indicates that there is not enough evidence to conclude that there are differences between sites. Table 16. Between site results for external reproducibility study. | Panel Member | Site | Neg. | Equivocal | Positive | Total | | --- | --- | --- | --- | --- | --- | | All | Between Site Overall - Final FISH Classification Number of Slides Across Lots/Runs/Readers | | | | | | | 1 | 41 | | 79 | 120 | | | 2 | 41 | | 79 | 120 | | | 3 | 45 | | 75 | 120 | | | Fisher-Freeman-Halton p-value: 0.8354 | | | | | | All | Between Site Overall - Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 80 | 94 | 66 | 240 | | | 2 | 83 | 1 | 156 | 240 | | | 3 | 87 | 83 | 70 | 240 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | | 1 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 0 | 30 | 10 | 40 | | | 2 | 0 | 1 | 39 | 40 | | | 3 | 2 | 33 | 5 | 40 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | | 2 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 0 | 18 | 22 | 40 | | | 2 | 0 | 0 | 40 | 40 | | | 3 | 0 | 2 | 38 | 40 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | | 3 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 0 | 12 | 28 | 40 | | | 2 | 0 | 0 | 40 | 40 | | | 3 | 3 | 26 | 11 | 40 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | PMA P110012: FDA Summary of Safety and Effectiveness Data {20} | Panel Member | Site | Neg. | Equivocal | Positive | Total | | --- | --- | --- | --- | --- | --- | | 4 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 40 | 0 | 0 | 40 | | | 2 | 40 | 0 | 0 | 40 | | | 3 | 40 | 0 | 0 | 40 | | | Fisher-Freeman-Halton p-value: 1.00 | | | | | | 5 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 40 | 0 | 0 | 40 | | | 2 | 40 | 0 | 0 | 40 | | | 3 | 39 | 1 | 0 | 40 | | | Fisher-Freeman-Halton p-value: 1.00 | | | | | | 6 | Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 0 | 34 | 6 | 40 | | | 2 | 3 | 0 | 37 | 40 | | | 3 | 3 | 21 | 16 | 40 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | A standard agreement analysis was provided comparing sites and demonstrated good agreement between sites. Table 17. Agreement Between Site A and Site B, Pair-wise | Panel | Site A | Site B | N | Called | | | | % Agreement | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | Pos by both readers | Neg by R1 & Pos by R2 | Pos by R1 & Neg by R2 | Neg by both readers | PPA | NPA | Overall | | All | 1 | 2 | 120 | 79 | 0 | 0 | 41 | 100.00 | 100.00 | 100.00 | | | 1 | 3 | 120 | 75 | 0 | 4 | 41 | 94.94 | 100.00 | 96.67 | | | 2 | 1 | 120 | 79 | 0 | 0 | 41 | 100.00 | 100.00 | 100.00 | | | 2 | 3 | 120 | 75 | 0 | 4 | 41 | 94.94 | 100.00 | 96.67 | | | 3 | 1 | 120 | 75 | 4 | 0 | 41 | 100.00 | 91.11 | 96.67 | | | 3 | 2 | 120 | 75 | 4 | 0 | 41 | 100.00 | 91.11 | 96.67 | R1 = Reader 1; R2 = Reader 2 Between-Lot Reproducibility – For the analyses on the final FISH classification, there was no significant association between the lots for the overall analysis (p-value of 0.9418) and by panel member (p-value ranged from 0.3220 to 1.00) by Fisher-Freeman-Halton (Table 18). A pair-wise agreement analysis was also provided (Table 19) which indicated good agreement between lots. Table 18. Between-lot results for external reproducibility study. | Panel Member | Lot | Neg. | Equivocal | Positive | Total | | --- | --- | --- | --- | --- | --- | | All | Between Lot Overall - Final FISH Classification Number of Slides Across Lots/Runs/Readers | | | | | | | 1 | 44 | | 76 | 120 | | | 2 | 41 | | 79 | 120 | | | 3 | 42 | | 78 | 120 | PMA P110012: FDA Summary of Safety and Effectiveness Data {21} | Panel Member | Lot | Neg. | Equivocal | Positive | Total | | --- | --- | --- | --- | --- | --- | | | Fisher-Freeman-Halton p-value: 0.9418 | | | | | | All | Probe Lot Overall - Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 85 | 88 | 67 | 240 | | | 2 | 82 | 47 | 111 | 240 | | | 3 | 83 | 43 | 114 | 240 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | | 1 | Probe Lot - Individual FISH Classification Number of Slides Across Lots/Runs/Readers/Enumerations | | | | | | | 1 | 1 | 32 | 7 | 40 | | | 2 | 0 | 15 | 25 | 40 | | | 3 | 1 | 17 | 22 | 40 | | | Fisher-Freeman-Halton p-value: <0.0001 | | | | | | 2 | 1 | 0 | 9 | 31 | 40 | | | 2 | 0 | 10 | 30 | 40 | | | 3 | 0 | 1 | 39 | 40 | | | Fisher-Freeman-Halton p-value: 0.0070 | | | | | | 3 | 1 | 3 | 21 | 16 | 40 | | | 2 | 0 | 6 | 34 | 40 | | | 3 | 0 | 11 | 29 | 40 | | | Fisher-Freeman-Halton p-value: 0.0001 | | | | | | 4 | 1 | 40 | 0 | 0 | 40 | | | 2 | 40 | 0 | 0 | 40 | | | 3 | 40 | 0 | 0 | 40 | | | Fisher-Freeman-Halton p-value: 1.00 | | | | | | 5 | 1 | 39 | 1 | 0 | 40 | | | 2 | 40 | 0 | 0 | 40 | | | 3 | 40 | 0 | 0 | 40 | | | Fisher-Freeman-Halton p-value: 1.00 | | | | | | 6 | 1 | 2 | 25 | 13 | 40 | | | 2 | 2 | 16 | 22 | 40 | | | 3 | 2 | 14 | 24 | 40 | | | Fisher-Freeman-Halton p-value: 0.0949 | | | | | Table 19. Agreement Between Lot A and Lot B, Pair-wise | Panel | Lot A (LA) | Lot B (LB) | N | Called | | | | % Agreement | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | Pos by both Lots | Neg by LA & Pos by LB | Pos by LA & Neg by LB | Neg by both Lots | PPA | NPA | Overall | | All | 2 | 3 | 120 | 78 | 0 | 1 | 41 | 98.73 | 100.00 | 99.17 | | | 2 | 1 | 120 | 76 | 0 | 3 | 41 | 96.20 | 100.00 | 97.50 | | | 3 | 2 | 120 | 78 | 1 | 0 | 41 | 100.00 | 97.62 | 99.17 | | | 3 | 1 | 120 | 76 | 0 | 2 | 42 | 97.44 | 100.00 | 98.33 | | | 1 | 2 | 120 | 76 | 3 | 0 | 41 | 100.00 | 93.18 | 97.50 | | | 1 | 3 | 120 | 76 | 2 | 0 | 42 | 100.00 | 95.45 | 98.33 | PMA P110012: FDA Summary of Safety and Effectiveness Data {22} Between-Reader Results – A pair-wise between-reader agreement analysis, stratified by site, was performed for is shown in Table 20. Table 20. Agreement Between Reader 1 and Reader 2 by site | Panel | Testing Site | N | Called | | | | % Agreement | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | Pos by both readers | Neg by R1 & Pos by R2 | Pos by R1 & Neg by R2 | Neg by both readers | NPA | PPA | Overall | | All | 1 | 120 | 79 | 0 | 0 | 41 | 100.00 | 100.00 | 100.00 | | | 2 | 120 | 77 | 1 | 2 | 40 | 97.56 | 97.47 | 97.50 | | | 3 | 120 | 69 | 0 | 10 | 41 | 100.00 | 87.34 | 91.67 | | | ALL | 360 | 225 | 1 | 12 | 122 | 99.19 | 94.94 | 96.39 | R1 = Reader 1; R2 = Reader 2 An analysis between readings for each panel member by reading the results over 2 lots, 5 runs, and 2 slides (20 readings) using the Gini Index is shown in Table 21. The Gini Index is a measure of the variability of repeated readings and is the probability that a pair of randomly selected readings will fall into different categories. Table 21. Gini Index analysis. | Panel Member | Total (n) | Gini Index | | | | --- | --- | --- | --- | --- | | | | Minimum | Median | Maximum | | 1 | 6 | 0.00 | 0.25 | 0.42 | | 2 | 6 | 0.00 | 0.09 | 0.50 | | 3 | 6 | 0.00 | 0.37 | 0.56 | | 4 | 6 | 0.00 | 0.00 | 0.00 | | 5 | 6 | 0.00 | 0.00 | 0.10 | | 6 | 6 | 0.10 | 0.26 | 0.58 | The percent agreement across all readers at all of the sites with the expected results is shown in Table 22 below. Table 22. Percent agreement between all readers with expected results. | | Expected Results | | | | | --- | --- | --- | --- | --- | | | | Positive | Negative | Total | | Observed | Positive | 463 | 0 | 463 | | | Negative | 17 | 240 | 257 | | | Total | 480 | 240 | 720 | PPA: 463/480 = 96.46% (94.40, 97.78) NPA: 240/240 = 100.00% (98.42, 100.00) OA: 703/720 = 97.64% (96.25, 98.52) Between-Reading – The end point for this analysis was the individual FISH classifications (positive/equivocal/negative) as determined by the scoring algorithm indicated in the package insert. A comparison of all of the readings of all the sections PMA P110012: FDA Summary of Safety and Effectiveness Data {23} of a tissue specimen for each panel member by reader over 2 lots, 5 runs, and 2 slides (20 readings) was calculated for inter-reader variability using the Gini Index. This analysis combines all panel members, while stratifying by site and reader. ## Precision (Repeatability): A precision analysis, the variance (standard deviation) and percent coefficient of variation (%CV) were estimated for within-reader component, between-reader component, between-run component, and between-site component for each lot and panel member. All the effects were considered as random for the analysis. The total lot variability was defined as the accumulation of within-reader component, between-reader component, between-run component, and between-site component. Overall precision by site, panel member, and by kit lot are provided in Tables 23-25. Some very high %CV values were noted; however this observation is not unexpected with FISH, particularly for specimens with low counts. In general, a higher degree of variance is to be expected in negative or low positive specimens due to the fewer signals. Table 23. Vysis ALK Break Apart FISH Probe Kit overall precision analysis by panel member. | Panel Member | n | Mean % Cells with ALK Rearrangement | Within-Reader Component | | Between-Reader Component | | Between-Run (Day) Component | | Between-Lot Component | | Between-Site Component | | Total | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | SD | %CV | SD | %CV | SD | %CV | SD | %CV | SD | %CV | SD | %CV | | 1 | 120 | 47.6 | 8.86 | 18.60 | 6.17 | 12.96 | 2.76 | 5.78 | 0.00 | 0.00 | 15.17 | 31.84 | 18.82 | 39.51 | | 2 | 120 | 61.5 | 7.91 | 12.87 | 6.80 | 11.06 | 0.00 | 0.00 | 0.00 | 0.00 | 10.00 | 16.27 | 14.46 | 23.51 | | 3 | 120 | 53.2 | 7.90 | 14.87 | 9.28 | 17.46 | 0.00 | 0.00 | 5.73 | 10.77 | 14.91 | 28.05 | 20.10 | 37.80 | | 4 | 120 | 1.3 | 1.52 | 118.17 | 0.84 | 65.19 | 0.00 | 0.00 | 0.00 | 0.00 | 0.79 | 61.93 | 1.91 | 148.50 | | 5 | 120 | 2.1 | 2.24 | 106.48 | 0.41 | 19.44 | 1.06 | 50.69 | 0.00 | 0.00 | 0.79 | 37.52 | 2.63 | 125.27 | | 6 | 120 | 46.2 | 15.15 | 32.76 | 0.00 | 0.00 | 9.07 | 19.62 | 1.81 | 3.92 | 10.24 | 22.14 | 20.49 | 44.32 | Table 24. Vysis ALK Break Apart FISH Probe Kit precision analysis by site. | Site | Panel Member | n | Mean % Cells with ALK Rearrangement | Within-Reader Component | | Between-Reader Component | | Between-Run (Day) Component | | Between-Lot Component | | Total | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | SD | %CV | SD | %CV | SD | %CV | SD | %CV | SD | %CV | | 1 | 1 | 40 | 47.5 | 6.68 | 14.06 | 0.00 | 0.00 | 2.41 | 5.06 | 0.00 | 0.00 | 7.10 | 14.94 | | | 2 | 40 | 50.3 | 6.67 | 13.28 | 3.81 | 7.58 | 2.85 | 5.67 | 0.00 | 0.00 | 8.19 | 16.30 | | | 3 | 40 | 54.0 | 6.04 | 11.20 | 1.67 | 3.10 | 1.97 | 3.65 | 0.00 | 0.00 | 6.57 | 12.18 | | | 4 | 40 | 1.6 | 1.48 | 92.70 | 0.00 | 0.00 | 0.74 | 46.35 | 0.00 | 0.00 | 1.66 | 103.64 | | | 5 | 40 | 2.2 | 1.97 | 91.85 | 0.32 | 14.71 | 0.91 | 42.25 | 0.00 | 0.00 | 2.20 | 102.17 | | | 6 | 40 | 38.7 | 10.39 | 26.85 | 0.00 | 0.00 | 10.28 | 26.57 | 0.00 | 0.00 | 14.62 | 37.77 | | 2 | 1 | 40 | 63.0 | 4.97 | 7.90 | 13.08 | 20.79 | 0.00 | 0.00 | 0.11 | 0.18 | 14.00 | 22.23 | | | 2 | 40 | 69.5 | 8.31 | 11.95 | 11.05 | 15.91 | 0.00 | 0.00 | 0.00 | 0.00 | 13.83 | 19.90 | | | 3 | 40 | 68.4 | 5.67 | 8.29 | 14.69 | 21.49 | 0.00 | 0.00 | 3.00 | 4.39 | 16.02 | 23.44 | | | 4 | 40 | 1.9 | 2.00 | 105.26 | 1.58 | 83.22 | 0.00 | 0.00 | 0.00 | 0.00 | 2.55 | 134.18 | | | 5 | 40 | 2.9 | 2.61 | 89.92 | 0.00 | 0.00 | 1.35 | 46.58 | 0.00 | 0.00 | 2.94 | 101.27 | | | 6 | 40 | 58.8 | 14.65 | 24.94 | 8.29 | 14.12 | 5.37 | 9.13 | 3.33 | 5.66 | 17.98 | 30.61 | | 3 | 1 | 40 | 32.5 | 12.89 | 39.72 | 0.00 | 0.00 | 9.37 | 28.88 | 0.00 | 0.00 | 15.93 | 49.10 | | | 2 | 40 | 64.7 | 8.63 | 13.33 | 1.41 | 2.19 | 4.79 | 7.40 | 0.00 | 0.00 | 9.97 | 15.40 | PMA P110012: FDA Summary of Safety and Effectiveness Data page 24 28 {24} PMA P110012: FDA Summary of Safety and Effectiveness Data page 25 | Site | Panel Member | n | Mean % Cells with ALK Rearrangement | Within-Reader Component | | Between-Reader Component | | Between-Run (Day) Component | | Between-Lot Component | | Total | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | SD | %CV | SD | %CV | SD | %CV | SD | %CV | SD | %CV | | | 3 | 40 | 37.2 | 10.90 | 29.30 | 6.32 | 17.00 | 6.04 | 16.22 | 9.61 | 25.83 | 16.96 | 45.58 | | | 4 | 40 | 0.4 | 0.84 | 239.05 | 0.00 | 0.00 | 0.34 | 95.83 | 0.00 | 0.00 | 0.90 | 257.54 | | | 5 | 40 | 1.3 | 2.07 | 165.89 | 0.71 | 56.57 | 0.87 | 69.28 | 0.00 | 0.00 | 2.36 | 188.47 | | | 6 | 40 | 41.3 | 19.12 | 46.35 | 0.00 | 0.00 | 10.60 | 25.71 | 3.02 | 7.31 | 22.07 | 53.50 | Table 25. Vysis ALK Break Apart FISH Probe Kit precision analysis by lot. | Lot | Panel Member | n | Mean % Cells with ALK Rearrangement | Within-Reader Component | | Between-Reader Component | | Between-Run (Day) Component | | Between-Site Component | | Total | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | | SD | %CV | SD | %CV | SD | %CV | SD | %CV | SD | %CV | | 1 | 1 | 40 | 40.8 | 11.23 | 27.53 | 0.00 | 0.00 | 5.58 | 13.67 | 8.16 | 20.01 | 14.97 | 36.68 | | | 2 | 40 | 57.5 | 7.94 | 13.83 | 4.37 | 7.61 | 0.00 | 0.00 | 9.04 | 15.73 | 12.80 | 22.28 | | | 3 | 40 | 41.9 | 8.38 | 20.00 | 7.65 | 18.25 | 0.00 | 0.00 | 16.96 | 40.47 | 20.40 | 48.69 | | | 4 | 40 | 0.8 | 1.10 | 136.93 | 0.00 | 0.00 | 0.64 | 80.28 | 0.63 | 78.81 | 1.42 | 177.22 | | | 5 | 40 | 1.8 | 1.84 | 102.44 | 1.18 | 65.73 | 1.13 | 62.73 | 0.00 | 0.00 | 2.46 | 136.93 | | | 6 | 40 | 40.9 | 15.27 | 37.37 | 0.00 | 0.00 | 12.07 | 29.54 | 3.68 | 9.00 | 19.81 | 48.48 | | 2 | 1 | 40 | 54.8 | 4.56 | 8.32 | 8.64 | 15.77 | 0.00 | 0.00 | 9.29 | 16.96 | 13.49 | 24.61 | | | 2 | 40 | 59.5 | 7.10 | 11.93 | 5.49 | 9.22 | 4.69 | 7.88 | 13.64 | 22.93 | 16.99 | 28.55 | | | 3 | 40 | 59.8 | 7.13 | 11.92 | 5.81 | 9.71 | 0.00 | 0.00 | 8.22 | 13.74 | 12.33 | 20.62 | | | 4 | 40 | 2.0 | 1.90 | 94.87 | 0.84 | 41.83 | 0.42 | 20.92 | 0.00 | 0.00 | 2.12 | 105.77 | | | 5 | 40 | 2.7 | 2.47 | 93.20 | 0.00 | 0.00 | 1.11 | 41.77 | 0.61 | 23.11 | 2.77 | 104.71 | | | 6 | 40 | 48.0 | 13.56 | 28.28 | 4.74 | 9.89 | 8.30 | 17.31 | 7.95 | 16.58 | 18.40 | 38.37 | | 3 | 1 | 40 | 47.3 | 9.40 | 19.88 | 9.34 | 19.74 | 4.13 | 8.73 | 23.84 | 50.40 | 27.58 | 58.32 | | | 2 | 40 | 67.5 | 8.63 | 12.78 | 9.46 | 14.02 | 0.00 | 0.00 | 0.00 | 0.00 | 12.80 | 18.97 | | | 3 | 40 | 57.8 | 8.15 | 14.10 | 12.90 | 22.31 | 0.00 | 0.00 | 18.18 | 31.46 | 23.74 | 41.07 | | | 4 | 40 | 1.1 | 1.45 | 138.01 | 1.34 | 127.78 | 0.00 | 0.00 | 0.88 | 83.57 | 2.16 | 205.81 | | | 5 | 40 | 1.9 | 2.35 | 126.77 | 0.00 | 0.00 | 0.95 | 51.28 | 0.63 | 34.19 | 2.61 | 140.96 | | | 6 | 40 | 49.9 | 16.47 | 33.00 | 0.00 | 0.00 | 5.69 | 11.41 | 17.86 | 35.79 | 24.95 | 50.00 | ## Conclusions Variability between readers was observed within and between sites, particularly prior to the resolution of the equivocal cases. Based on the Gini Index results, readers at the same site seemed to perform more similarly than between sites. Site 3 seemed to show the greatest differences between readers which could not be fully accounted for and specimens (panels) one and three showed more variability in results than the other four specimens from the external reproducibility study. This could result from the observation that if the readers do not enumerate the same areas and/or cells, variability in results is expected. Overall the reproducibility between readers, sites, tissue specimens, and the ProbeChek ALK Control slides was determined to be good. ## B. Animal Studies None {25} PMA P110012: FDA Summary of Safety and Effectiveness Data page 26 # C. Additional Studies ## Assay Run Validity Rate – Failure Rate The assay run informative rate, defined as the number of passing slides divided by the number of total slides multiplied by 100, was calculated from the reproducibility studies. Based on the results of the ProbeChek ALK Negative and Positive Control slides, the assay run failure rate was calculated to be 3.08% and the passing rate was calculated to be 96.92%. In a total of four invalid assay runs, both the ProbeChek ALK Negative and Positive Control slides were invalid resulting in a total frequency of 8 failures. Table 26. ProbeChek ALK Control Slides - Pass and Fail rate. | ProbeChek ALK Control slides | Disposition | Frequency | Rate (%) | | --- | --- | --- | --- | | Positive and Negative slides | Fail | 8 | 3.08 | | | Pass | 252 | 96.92 | | Positive slides | Fail | 4 | 3.08 | | | Pass | 126 | 96.92 | | Negative slides | Fail | 4 | 3.08 | | | Pass | 126 | 96.92 | ## Rates of Information and Uninformative Slides The rates of informative (successful hybridization) and uninformative (failed hybridization) slides were calculated from the reproducibility studies which used primarily non-clinical trial specimens. The rate of uninformative slides was defined as the number of uninformative slides divided by the number of total slides multiplied by 100. The rate of informative slides was calculated to be 95.24% and the rate of uninformative slides was calculated to be 4.76%. Table 27. External reproducibility study uninformative slide rate. | Disposition | Frequency | Rate (%) | | --- | --- | --- | | Informative | 720 | 95.24 | | Uninformative | 36 | 4.76 | | Total | 756 | 100.00 | In order to collect the planned sample size of 390 samples for the Concordance study (discussed in Sect. III and not used to support the PMA), Abbott Molecular reviewed 425 slides from the 1005 trial. The uninformative rate of the Vysis ALK Break Apart FISH Probe Kit for this subset of samples was 8.47%. The results are shown in Table 28. Table 28. Uninformative rate with a subset of 1005 clinical trial specimens. | Disposition | Frequency | Rate (%) | | --- | --- | --- | | Informative | 389* | 91.53 | | Uninformative | 36 | 8.47 | | Total | 425 | 100.00 | * One slide was mistakenly included in the randomization as a negative specimen however the actual result was uninformative. {26} Some observed causes for uninformative results are listed below: - Control slide(s) uninformative or failed - Less than 50 nuclei are evaluable - Fluorescent haze/glow covering nuclei - No FISH signals - Weak FISH signals - Debris, bacteria, non-specific signals or high background - Borders of nuclei are overlapping or not distinguishable - Poor nuclear integrity - Wrong specimen type The uninformative rate with the clinical specimens was nearly twice that as seen with the reproducibility study specimens. The precise reason for this difference is not clear as the participating sites from the external reproducibility study are three of the four clinical trial testing sites and the only difference are the specimens which were used in the two studies. If the cause for the higher uninformative rate is due to differences in specimens, then the rate seen with the clinical trial specimens may be more representative. ## Vysis ALK Break Apart FISH Probe Kit Shelf life and Stability Real-time kit stability was assessed under the intended storage conditions (ISC), inverted storage (INVERT), and combined transport and temperature extremes conditions (TTE) for the Vysis ALK Break Apart FISH Probe Kit and DAPI I Counterstain to provide evidence of ruggedness to various stresses for the material. The expiration date assigned for the kit is based on the expiration date of the shortest dated component. Stability was measured at 0, 3, 6, 9, and 12 months for ISC conditions and 0, 6, and 12 months for TTE/INVERT conditions. The zero (0) month time point for ISC storage condition served as the baseline for both the ISC and TTE/INVERT conditions. Three lots of the Vysis ALK Break Apart FISH Probe Kits were stored upright at -20°C while protected from light for the duration of the stability study. To evaluate temperature and transport extremes and inverted storage (TTE/INVERT), one lot of the Vysis ALK Break Apart FISH Probe Kit was removed from storage and subject to the conditions below: - Removed from -20°C freezer - 1 cycle of dry ice for 48 hours ±2 hours - 1 cycle of 25°C for 72 hours ±2 hours - 1 cycle of 40°C for 72 hours ±2 hours - 1 cycle of -20°C for 24 hours ±2 hours - 3 freeze/thaw cycles** - Return to ISC - Store inverted (INVERT) PMA P110012: FDA Summary of Safety and Effectiveness Data {27} For the freeze/thaw cycles, the kits for TTE/INVERT testing were removed from $-20^{\circ}\mathrm{C}$ freezer and left at room temperature for $\leq 8$ hours at room temperature until thawed, protected from the light, then returned to $-20^{\circ}\mathrm{C}$ for $\geq 16$ hours. 3 cycles of freeze/thaw were performed for each TTE/INVERT simulation. At the time of approval the Vysis ALK Break Apart FISH Probe Kit has been tested at the 0 (Baseline), 3, 6, 9, and 12 months and expiration dating is valid for 12 months from the date of manufacture. Stability studies are ongoing to support 24 months. ## ProbeChek ALK Negative and Positive Control Shelf life Stability Real-time studies to establish shelf life integrity and expiration dating for the ProbeChek ALK Negative and Positive Control slides will be conducted in accordance with the ISC and TTE testing schedule described above. Three lots each of the control slides were stored upright at $15 - 30^{\circ}\mathrm{C}$ (ISC) for the duration of the stability study and one lot of each control slide tested according to the following parameters: - Removed from ISC - 1 cycle of $-20^{\circ}\mathrm{C}$ for 72 hours $\pm 2$ hours - Return to ISC for 24 hours - 1 cycle of $45^{\circ}\mathrm{C}$ for 72 hours $\pm 2$ hours - Return to ISC Data was provided for only the baseline time point. Because the ProbeChek ALK Control slide lots ranged from 4-8 months from the original manufacturing date at baseline testing and data from an additional functional testing study will allow for initial expiration dating for six months. A real-time stability protocol was submitted and approved to allow for the establishment the stability dating to 24 months from the date of manufacture. ## Paraffin Pretreatment IV &amp; Post-Hybridization Wash Buffer Kit Shelf life Stability Real-time studies to establish shelf life integrity and expiration dating for the Vysis Paraffin Pretreatment IV &amp; Post-Hybridization Wash Buffer Kit will be conducted under four conditions. The ISC condition (2-8°C, upright) and an alternative ISC condition [in accordance with the conditions recommend to the customer after receipt, where the pepsin is separately stored at -20°C (±10°C)] will be tested on 3 separate kit lots and tested at 0 month (baseline) and 6, 13, 18, 24 and 25 (final time point) month time points. TTE/INVERT (combined) and 15-30°C (upright) testing will be conducted on a single kit lots at 0 (baseline), 13, 18 and 25 month time points. The TTE/INVERT condition according to the following parameters: - Removed from $2 - 8^{\circ}\mathrm{C}$ storage, - 1 cycle of $15 - 30^{\circ}\mathrm{C}$ for 168 hours $\pm 2$ hours - 1 cycle of $25^{\circ}\mathrm{C}$ for 72 hours $\pm 2$ hours - 1 cycle of $40^{\circ}\mathrm{C}$ for 24 hours $\pm 2$ hours - 1 cycle of $-20^{\circ}\mathrm{C}$ for 24 hours $\pm 2$ hours PMA P110012: FDA Summary of Safety and Effectiveness Data {28} - Return to 2-8°C - Store inverted (INVERT) Data was provided for only the baseline time point. Because one kit was at 1.5 months after kit manufacture at baseline testing and dat…