Virtual Library

Start Your Search

Peter B Illei



Author of

  • +

    P1.01 - Advanced NSCLC (ID 757)

    • Event: WCLC 2017
    • Type: Poster Session with Presenters Present
    • Track: Advanced NSCLC
    • Presentations: 2
    • +

      P1.01-007 - ALK Testing Trends and Patterns Among Community Practices in the United States (ID 8654)

      09:30 - 16:00  |  Presenting Author(s): Peter B Illei

      • Abstract

      Background:
      The CAP/IASLC/AMP molecular testing guidelines recommend ALK testing on patients with lung adenocarcinoma, regardless of clinical characteristics. FISH is the recommended assay to detect ALK rearrangement, however other assays, such as NGS and IHC, are available. There have been limited published data to assess adherence to ALK testing guidelines using large real-world data sources. The objective of this study was to assess real-world ALK testing patterns among community practices in the United States.

      Method:
      The Flatiron database provides real-world clinical data collected from EHRs used by US cancer care providers. The Flatiron network comprises ~15% of US cancer patients and is geographically and demographically diverse. Patients with ≥2 visits within the Flatiron Network after Jan 1, 2011, >=18 years of age, and an stage IIIB/IV NSCLC diagnosis from 2011 through 2017 Q1 were included in this analysis. Logistic regression was used to identify patient characteristics associated with receiving ALK testing.

      Result:
      Of 29,903 patients identified from community-based clinics (mean age: 71.6, 52.2% male), ALK testing rates have steadily increased over time from 32.2% in 2011 to 61.0% in 2016 for all NSCLC patients, and 41.0% in 2011 to 74.0% in non-squamous patients. Patients that are younger, no history of smoking, women and living in the West region were more likely to be tested for ALK. Patients with Medicaid insurance, recurrent disease and squamous histology were less likely to be tested. The most common first assay to test for ALK was FISH (70%) followed by NGS (8%), PCR (4%) and IHC (1%). The median time from specimen receipt by lab to test result ranged from 6 days (FISH) to 11 days (NGS). Patients who had NGS testing were more likely to initiate chemotherapy prior to test result (34% of patients tested with NGS) than FISH (20%). 1235 patients had at least one FISH and another ALK test, with the percent agreement between FISH and other assays (NGS, PCR, IHC) ranging from 92% to 97%.

      Conclusion:
      Several patient characteristics predicted ALK testing indicating that some subgroups of patients may be under tested, according to guidelines. Consistent with guidelines, FISH was the most common assay and turnaround times from lab receipt to test result was under 2 weeks. There was a high agreement between FISH and NGS, indicating the potentially clinical utility of NGS, however NGS had also the longest turn around time and the highest proportion of patients initiating treatment prior to test results.

    • +

      P1.01-008 - Real-World Patient Characteristics, Testing and Treatment Patterns of ALK+ NSCLC (ID 8681)

      09:30 - 16:00  |  Author(s): Peter B Illei

      • Abstract

      Background:
      Based on clinical trials, ALK+ patients have been described as typically younger and never/former smokers, however patients enrolled in clinical trials may be different than those in the real-world. While the ALK positivity rate has been described as about 4% among all NSCLC patients, limited information is available on the positivity rates in patient subgroups. The objective of this study is to describe the real-world ALK positivity rates, patient characteristics and treatment patterns in ALK+ NSCLC patients.

      Method:
      The Flatiron database provides real-world clinical data collected from EHRs used by US cancer care providers. The Flatiron network comprises ~15% of US cancer patients and is geographically and demographically diverse. Patients with ≥2 visits within the Flatiron Network after Jan 1, 2011, ≥18 years of age, ≥1 ALK+ test result and an stage IIIB/IV NSCLC diagnosis from 2011 through 2017 Q1 were included in this analysis. Logistic regression was used to examine the association of ALK positivity and initiation of ALK inhibitor therapy based on patient characteristics. Survival model adjusting for censoring was used to estimate the time to ALK inhibitor order.

      Result:
      599 out of 15,551 ALK tested patients were identified to have an ALK positive test result, for a positivity rate of 3.9%. The ALK positivity rate varied by age (<65: 6.3% vs. ≥65: 2.9%), smoking status (no history of smoking: 11.6% vs. history of smoking: 2.3%), and histology (non-squamous: 4.0% vs. squamous: 1.8%). Factors associated with ALK positivity included younger age, academic practice, male, non-squamous histology, and no history of smoking. 78% of patients with ALK+ disease had evidence of an order for an ALK inhibitor after NSCLC diagnosis. The median time from test result to ALK inhibitor order was 24 days, with 42% of patients without an order for an ALK inhibitor within 90 days. Among patients with an order for an ALK inhibitor, 23% received chemotherapy prior to their ALK test result and 20% received chemotherapy after their test result but before the first order of ALK inhibitor. Patients diagnosed after 2014 and patients who received chemotherapy prior to the ALK test result were more likely to have an order for an ALK inhibitor.

      Conclusion:
      The ALK positivity rate and patient characteristics in this real-world NSCLC population are consistent with clinical trials, with some subgroups having higher positivity rates. ALK inhibitors were the most frequently ordered treatment, however many patients had a delayed time to ordering the ALK inhibitor.

  • +

    P2.02 - Biology/Pathology (ID 616)

    • Event: WCLC 2017
    • Type: Poster Session with Presenters Present
    • Track: Biology/Pathology
    • Presentations: 2
    • +

      P2.02-044 - Pulmonary Findings in 7 Autopsy Cases of Patients Treated with Immune Checkpoint Inhibitors (ID 10263)

      09:30 - 16:00  |  Presenting Author(s): Peter B Illei

      • Abstract

      Background:
      Side effects of immune checkpoint therapy are milder than with standard chemotherapy. Pulmonary toxicity has been described in 5% of patients, most of which are low grade pneumonitis with varied radiologic and pathologic appearances. Here we report lung pathology findings in 7 autopsy cases of patients who died while on immune checkpoint therapy.

      Method:
      Patient characteristics are shown in table 1. We evaluated tumor burden, the presence of tumor infiltrating lymphocytes (TIL), and type of lung injury. TIL were quantitated using a 4 tier system (0 -3+ = none, mild, moderate, extensive)

      Table1. Summary of demographic and clinical information
      Case Age Sex Immune Tx Pneumonitis Chest Imaging
      1 38 WM 3 mo Yes GGO, multifocal
      2 65 WM 9 mo No Bilateral pleural effusion
      3 54 WF single dose Yes Bilateral reticular opacities
      4 30 BM 4 mo No GGO, patchy
      5 59 WM 7 mo No Pulmonaryemboli and tumor
      6 72 WM 25 mo Yes GGO, bilateral, diffuse
      7 68 WM 5 mo Yes GGO, bilateral, diffuse


      Result:
      Residual viable tumor with little or no therapy effects was present in all 7 cases. The 4 cases where pneumonitis was diagnosed clinically had at least focal DAD. The majority of tumors had at least scattered TIL present, while one tumor had a large number TIL.
      Table 2. Summary of autopsy findings
      Case Tumor TIL LVI DAD Pneumonia Tumor necrosis
      1 ACA in 5 lobes 1+ Yes* Focal No <5%
      2 MM in LLL, LUL 3+ No Focal RLL No None
      3 ACA in 4 lobes 1+ Yes LUL, RUL, LLL LLL, RLL, LUL None
      4 MM in 4 lobes 1+ Yes No No None
      5 MM in RUL 1+ No No No None
      6 SCC in RLL, LL, LUL 2+ Yes RUL, RML, RLL No 10% and fibrosis
      7 SCC in 5 lobes 1+ Yes LUL, LLL No 20-30%
      ACA: adenocarcinoma; MM: Malignant melanoma; SCC: squamous cell carcinoma; *extensive LVI

      Conclusion:
      The majority of patients with clinical diagnosis of pneumonitis had DAD at autopsy and variable amount of viable tumor in the lungs and at least a few tumor infiltrating lymphocytes. Additional studies are pending to further characterize the phenotype of the TIL and to determine PD1 and PDL1/PDL2 expression on the tumor cells.

    • +

      P2.02-045 - PD-L1 Assessment in Cytology Samples (ID 10485)

      09:30 - 16:00  |  Presenting Author(s): Peter B Illei

      • Abstract

      Background:
      Pembrolizumab therapy for non-small cell lung cancer requires patient selection using PD-L1 immunohistochemistry (IHC). FDA approval was based on staining of resections or core biopsies. There is only limited data on PD-L1 expression in fine needle aspiration (FNA) specimens including EBUS of mediastinal lymph nodes.

      Method:
      Immunohistochemistry (IHC) was performed on an automated platform (Ventana Benchmark Ultra) using clone 22C3 (Dako) and the Optiview detection system (Ventana Medical Systems) on formalin fixed paraffin embedded FNA cell blocks (n:49), cell blocks from aspirated fluids (n:14), core biopsies (n:21) and 4 transbronchial biopsies. The cohort included of 18 squamous cell carcinomas, 56 adenocarcinomas (ACA), 2 adenosquamous carcinoma, 1 NUT-1 carcinoma, 7 NSCLC and 2 melanomas (age: 24-89, mean: 66, median: 69; 43 female and 45 male). Membranous PD-L1 staining of any intensity and extent was recorded in at least 100 tumor cells (tumor proportion score). The tumors were grouped as: no staining (< 1%), low expression (1-49%) and high expression (50% or more).

      Result:
      Twenty (23.3%) tumors showed no staining, 26 (30.2%) low expression and 40 (46.5%) high expression. Of the 56 adenocarcinomas 21 (37.5%) showed high expression, while of the 18 squamous cell carcinomas 7 (39%) showed high PD-L1 expression. The other tumors with high expression included one adenosqaumous carcinoma and 11 poorly differentiated carcinomas. One EBUS biopsy also had PDL1 assessed on a transbronchial biopsy of the primary tumor showing similar staining (30% versus 20%).

      Conclusion:
      PD-L1 immunohistochemistry appears to be feasible using formalin fixed cell blocks of fine needle aspirates and aspirated fluid specimens containing adequate number of viable tumor cells. High PD-L1 expression was seen in approximately half the tumors, which is greater than has been observed in resections/core biopsies, this finding merits further study. High expression of PD-L1 was seen in both squamous cell carcinoma and adenocarcinomas.