Search

2602 Background: A CLIA-certified organoid based drug sensitivity assay (a cancer organogram) has been developed for all solid tumors. An actionable report of organogram sensitivities to endocrine, chemotherapy and targeted agents, produced a drug sensitivity score as a tool to inform therapy decision making. Objectives: To evaluate the success rate of organoid derivation, the organogram drug responses across cancer types and to analyze the impact of the organogram report on therapeutic decision making. Methods: From 2016 to 2020, 628 cancer organograms were performed, with 513 tumor samples from 419 cancer patients. Within 48 hours of collection, fresh samples of tumor cells obtained from core biopsies, surgical excisions, or fluids were cultured, the majority as 3D organoids. Drug screens were performed with a library of up to 220 drugs, and dose-response was evaluated across a range of concentrations for each drug. Organogram sensitivity was ranked as response in five categories based on SPM Score: Exceptional (SPM15/14), Good (13/12), Moderate to Low (11/9), and None( < 9). 118 drugs on average were tested per screen (range: 68-152), so in a total of 628 organograms, more than 70,000 individual drug trials have been performed. The median turnaround time was 28 days (range: 19.5-51.5). Results: Of the 513 collected samples, 314 were fresh specimens: 96 core biopsies, 151 surgical specimens, and 67 fluids (pleural effusions or ascites), with an organoid derivation success rate of 58.3%, 78%, and 88%, respectively. Overall success rate in organoid derivation was 70.2%. Samples with poor viability and low tumor cell count (22%) were rejected. The primary cancer types tested were ovarian (n = 92, 17.9%), breast (n = 73, 14.2%), colorectal (n = 70, 13.6%), pancreatic (n = 51, 9.9%), cholangiocarcinoma (n = 42, 8.1%), and other solid tumors (n = 185, 36%). Median age of patients was 56 years old (range: 5-83), most of them heavily pretreated. 20.45% of drugs screened had exceptional and good responses (SPM score 15-12) (SD: 17.92%). We reviewed genomic data from 374 third-party genomic reports. The most frequent genomic alterations found were TP53 (n = 143, 38.2%), BRCA1 and BRCA2 (n = 47, 12.5%) CDKN2A (n = 42, 11.2%), FGFR1/2/3/4 (n = 41, 10.9%), and PIK3CA (n = 38, 10.1%). Post-test treatment information is available for a subset of 61 patients. The treating physician made an organogram-guided therapeutic decision in 32/44 patients with post test treatment drugs scored (72%). Conclusions: The cancer organogram test has a high rate of success in generating an actionable report that identifies therapies for patients with limited therapeutic options, including those with no known genomic biomarkers. The organogram guided selection of therapeutics for a significant subset of patients, nearly 4 times the rate reported with genomic testing alone.

e14032 Background: Anti-PD-(L)1 checkpoint inhibitors (CPI) are remarkable due to durable responses, tolerability, and improvements in survival across multiple solid tumors. Unfortunately, CPI produce objective responses in only a small subset of solid tumors patients (pts). We hypothesized that PD-L1 and IRF-1 expression on CTC may correlate with CPI response, and evaluate their expression in this pilot study. Methods: Pts with metastatic solid tumors initiating anti-PD(L)-1 mono- or combination therapy were eligible. Pts had 7.5 cc of blood drawn for CTC evaluation at 3 timepoints: baseline, after 1 CPI dose, and 3-6 months after starting CPI or at time of CPI discontinuation. Peripheral blood CTC were isolated with the AccuCyte kit and stained for CK/EpCAM, anti-CD45, nuclear dye, PD-L1, and IRF-1. CTCs were then identified with CyteFinder software. Results: 16 pts have enrolled to date with median age 71 years (range 28-83) and median ECOG = 1 (range 0-3). 50% were female. Diagnoses include NSCLC (50%), head and neck cancer (13%), breast cancer (13%), and others (24%). CTC were identified at baseline in 13/16 pts (# CTCs range 0-101). 12 pts with baseline CTC have had restaging scans since starting CPI. By RECIST 1.1, 3 pts had partial response (PR), 1 had stable disease (SD), and 8 had progressive disease (PD). The pts with PR more frequently had IRF-1 staining on baseline CTC (Table), and 1/3 had PD-L1 staining. 6 pts had tumor PD-L1 staining available for correlation, and 3/6 of those pts were PD-L1+ (tissue PD-L1 ≥ 1%). 1/3 pts with PD-L1+ tumor staining had CTCs that were PD-L1+. In the 3 pts with negative PD-L1 tumor staining, 1 pt had a mixed PD-L1+/PD-L1- CTC population and the others had no PD-L1 staining. Conclusions: CTC isolation and enumeration seem feasible across tumor types. Pts with PR to anti-PD-(L)1/ might have higher prevalence of IRF-1+ CTC; however, results are preliminary and more pts and samples are being evaluated. [Table: see text]

e14281 Background: Trastuzumab is an important drug in the treatment (tx) of HER2+ breast cancer, but not all tumors are responsive. Immune-enriched tumors have a better response to tx, and a biomarker to identify these tumors would be valuable. Multiparametric functional MRI is a non-invasive technique that can assess the entire tumor without the sampling error of biopsy. In this pilot study, we evaluate breast MRI as a potential biomarker for tumor immune response to trastuzumab. Methods: Women with de novo cT1-2N0 HER2+ breast cancer were enrolled and received paired multiparametric breast MRI examinations (including DWI and DCE sequences) prior to and ~2 weeks after a single dose of trastuzumab, followed by surgery or biopsy (if additional neoadjuvant tx planned). Tumor peak signal enhancement ratio (SER) and mean apparent diffusion coefficient (ADC), reflecting tissue perfusion and tissue cellularity, were determined from the images. Pre- and post-tx tissue were immunologically profiled with TIL counts and an immune-focused RNA gene expression panel. SER and ADC were correlated with TIL and RNA gene expression using Spearman’s rho. Results: To date, 8 women have enrolled and 6 had pre- and post-tx tissue available. The primary endpoint was correlation between SER and ADC measurements with tumor TILs. Secondary endpoints included correlating SER with angiogenesis and inflammation markers and correlating changes in SER and ADC with apoptosis/cell death markers. About half of the patients had TIL-enriched (TIL≥ 50%) tumors at baseline, and after trastuzumab the majority of tumors remained enriched or had an increase in TILs. Results suggest that pre-tx ADC had a negative correlation with TILs by histology and gene expression, although correlations did not reach statistical significance in this small cohort. Results also suggest a strong correlation between pre-tx SER and VEGF-A (r = 0.83) with a trend towards significance (P = 0.10). A decrease in SER post-tx had a strong correlation with an increase in caspase 3 expression (r = -0.85, p = 0.05), which suggested tx effect. Conclusions: Functional MRI may provide a novel biomarker of breast tumor immune response to trastuzumab. Results from additional patients will be available at the meeting.