PURPOSE Current clinical risk stratification methods for localized prostate cancer are suboptimal, leading to over- and undertreatment. Recently, machine learning approaches using digital histopathology have shown superior prognostic ability in phase III trials. This study aims to develop a clinically usable risk grouping system using multimodal artificial intelligence (MMAI) models that outperform current National Comprehensive Cancer Network (NCCN) risk groups. MATERIALS AND METHODS The cohort comprised 9,787 patients with localized prostate cancer from eight NRG Oncology randomized phase III trials, treated with radiation therapy, androgen deprivation therapy, and/or chemotherapy. Locked MMAI models, which used digital histopathology images and clinical data, were applied to each patient. Expert consensus on cut points defined low-, intermediate-, and high-risk groups on the basis of 10-year distant metastasis rates of 3% and 10%, respectively. The MMAI's reclassification and prognostic performance were compared with the three-tier NCCN risk groups. RESULTS The median follow-up for censored patients was 7.9 years. According to NCCN risk categories, 30.4% of patients were low-risk, 25.5% intermediate-risk, and 44.1% high-risk. The MMAI risk classification identified 43.5% of patients as low-risk, 34.6% as intermediate-risk, and 21.8% as high-risk. MMAI reclassified 1,039 (42.0%) patients initially categorized by NCCN. Despite the MMAI low-risk group being larger than the NCCN low-risk group, the 10-year metastasis risks were comparable: 1.7% (95% CI, 0.2 to 3.2) for NCCN and 3.2% (95% CI, 1.7 to 4.7) for MMAI. The overall 10-year metastasis risk for NCCN high-risk patients was 16.6%, with MMAI further stratifying this group into low-, intermediate-, and high-risk, showing metastasis rates of 3.4%, 8.2%, and 26.3%, respectively. CONCLUSION The MMAI risk grouping system expands the population of men identified as having low metastatic risk and accurately pinpoints a high-risk subset with elevated metastasis rates. This approach aims to prevent both overtreatment and undertreatment in localized prostate cancer, facilitating shared decision making.

In pre-treatment biopsy samples from a randomized Phase 3 trial cohort, GC demonstrated its ability to further risk stratify clinically high-risk men demonstrating an independent association of GC score with DM and MFS. High-risk prostate cancer is a heterogeneous disease state and GC can improve risk stratification to help personalize shared decision-making. NRG-GU009/PREDICT-RT (NCT04513717) aims to determine the optimal therapy based on GC score for high-risk prostate cancer.

4523 Background: Chemoradiation therapy (CRT) is an organ conserving approach in the treatment of locally advanced bladder cancer. Chemoradiation is thought to potentially result in immunogenic stimulation, and bladder cancer is often a tumor with high immune cell infiltration. Thus, we aimed to profile the tumor immune microenvironment of bladder cancer and identify prognostic immune biomarkers for CRT response by profiling tumor samples from NRG/RTOG 0524 and 0712, two prospective trials of CRT in muscle invasive bladder cancer (MIBC). Methods: Pretreatment tissue samples from both trials were profiled using Cofactor Genomics ImmunoPrism, an RNA sequencing assay that uses gene expression profiles to quantify immune cell populations in the tumor microenvironment (TME). Differential gene expression was estimated for different immune cell type proportions across samples. Kaplan-Meier survival analysis and log rank tests were performed to evaluate differences in overall survival (OS) stratified by genes influenced by immune cell proportions or genes associated with immune response signatures. Results: A total of 70 samples (43 from RTOG 0524 and 27 from RTOG 0712) underwent analysis using the ImmunoPrism assay. Immune cell proportions were as follows: CD8 T cells: median 1.2%, CD4 T cells: median 0.8%, Treg cells: median 9.2%, CD19 B cells: median 5.1%, M2 macrophages: median 0.8%, M1 macrophages: median 0%. Unbiased clustering based on gene expression profiles driven by immune cell proportions demonstrated two groups: cluster 1 with a low percentage of immune cells and shorter OS (median 31 months) and cluster 2 with a high percentage of immune cells and longer OS (median 101 months, p = 0.036). Higher expression of genes associated with T cell infiltration ( CD8A and ICOS) was associated with improved OS (104 vs 35 months, p = 0.028, HR = 0.48 (0.25 – 0.94), p = 0.031) as was higher expression of IDO1, which is associated with the interferon gamma pathway (104 vs 35 months, p = 0.042, HR = 0.49 (0.24 – 0.99), p = 0.046). Conclusions: Bladder tumors have a wide range of immune cell infiltration in the TME. Increased immune cell proportions are prognostic for OS following CRT, as well as a higher expression of genes associated with T cell infiltration interferon gamma signaling. These findings have implications for the integration of immunotherapy in the definitive management of MIBC; and can be explored further in the ongoing NRG/SWOG 1806 trial.

Purpose/Objective(s)The use of radiation therapy (RT) and androgen deprivation therapy (ADT) following radical prostatectomy (RP) has been recently addressed by the NRG/RTOG 0534 phase III trial, which compared prostate bed RT alone (arm 1), prostate bed RT + short term (st-ADT) (arm 2), and prostate bed RT + st-ADT + pelvic lymph node (LN) RT (arm 3). A wide range of LN volumes were removed at RP prior to enrollment. We sought to study the influence of LN dissection volumes on clinical outcomes.Materials/MethodsOur hypothesis was that the number of LNs resected at RP on NRG/RTOG 0534 influences the benefit of pelvic LN RT (arm 3). Outcomes of interest included freedom from progression (FFP), time to distant metastasis (DM) and toxicity. Pretreatment characteristics were compared across quartiles of number of LNs dissected using the Kruskal Wallis test for categorical variables and the f-test from analysis of variance for continuous variables. Adjusted multivariable (MV) Cox proportional hazards models (cause-specific in the presence of competing risks) were used to assess the interaction between treatment arm and LNs dissected (0-1 vs. 2+ and as a continuous variable) with FFP and time to DM.Results1,626 eligible patients had data on number of LNs dissected, 538 in arm 1, 543 in arm 2, and 545 in arm 3. 606 patients had 0 LNs removed, 228 had 1-2, 414 had 3-7, and 378 had over 7 LNs removed (max of 54). A higher number of LNs removed was associated with more adverse Gleason grade (p< 0.01), T-stage (p< 0.01), seminal vesicle invasion (p < 0.01), and academic medical center enrollment (p < 0.01). Controlling for differences in patient characteristics and continuous LN dissected, patients on arm 3 (vs. arm 2) did not demonstrate improvement in FFP (HR 1.19, 95% CI: 0.93-1.51, p = 0.2) nor was the interaction between treatment and LN dissected significant. Upon MV analysis for DM, there was a significant interaction between continuous number of LNs dissected and treatment arm. Specifically, patients with 0 LNs removed on arm 3 had a significant improvement in DM when compared with arms 1 and 2 (vs. arm 1 HR 2.49, 95% CI: 1.46-4.26; vs. arm 2 HR 1.75, 95% CI: 1.02-3.01). Patients with 1 node removed and on arm 3 had a significant improvement in DM when compared with arms 1 and 2 (vs. arm 1 HR 2.34, 95% CI: 1.42-3.85; vs. arm 2 HR 1.67, 95% CI: 1.00-2.77). The benefit of arm 3 vs. arm 2 was no longer significant in patients with 2+ LNs removed as the 95% CIs crossed 1.0. 2+ vs. 0-1 pelvic LNs removed and treatment on arm 3 vs. 1 exhibited higher rates of any grade renal/genitourinary toxicity (HR=1.16, 95% CI:1.02-1.31; HR=0.86, 95% CI: 0.75-0.99, respectively).ConclusionThe extent of surgical pelvic LN dissections influences the benefit of post-operative pelvic LN RT on FFP and DM. Patients with 0-1 LNs removed, and randomized to pelvic LN RT and st-ADT, benefitted significantly from the addition of pelvic LN RT. Although no improvements with pelvic LN RT were seen in patients with 2+ nodes removed, these analyses had low statistical power.

Limited stage small cell lung cancer (LS-SCLC) is treated with standard of care platinum/etoposide (EP) and thoracic radiation therapy (TRT) with curative intent, however the majority of patients are not cured and median overall survival is approximately 30 months. Addition of atezolizumab to chemotherapy in extensive stage SCLC has improved progression free and overall survival in a non-curative setting leading to hope that addition of an immune checkpoint inhibitor to standard chemoradiotherapy could benefit LS-SCLC patients.

This project was supported by grants U10CA180868 (NRG Oncology Operations), U10CA180822 (NRG Oncology SDMC), U24CA180803 (IROC) from the National Cancer Institute (NCI) and Genentech.

This is the first validation of any gene expression biomarker on pre-treatment biopsy samples from prospective randomized trials and demonstrates an independent association of GC score with DM, PCSM, and OS. High-risk prostate cancer is a heterogeneous disease state and GC can improve risk stratification to help personalize shared decision-making. NRG-GU009/PREDICT-RT will further determine the optimal therapy based on GC score. NCT04513717.

Clinical outcomes for limited stage small cell lung cancer (LS-SCLC) remain suboptimal. Standard of care chemoradiation with platinum/etoposide and thoracic radiation to 45 Gy delivered twice daily followed by prophylactic cranial irradiation yields a median overall survival of 30 months. LU005 is a randomized phase II/III trial designed to test the addition of atezolizumab to concurrent chemoradiation (ClinicalTrials.gov Identifier: NCT03811002). Patients with LS-SCLC (Tx-T4, N0-N3, M0) are randomly assigned in a 1:1 ratio to either standard chemoradiation, consisting of thoracic radiation (45 Gy twice daily or 66 Gy daily) with concurrent platinum/etoposide chemotherapy, or the experimental arm, consisting of the same chemoradiation regimen plus the addition of atezolizumab beginning concurrently with thoracic radiation, and continued every 3 weeks for 12 months duration. Thoracic radiation begins with the second cycle of chemotherapy in both treatment arms. Stratification variables include radiation schedule (once daily vs. twice daily), chemotherapy (cisplatin vs. carboplatin), sex, and performance status (PS 0/1 vs. 2). Prophylactic cranial radiation is recommended for patients who have a response to treatment. The phase II primary endpoint is progression free survival (PFS) and the phase III primary endpoint is overall survival (OS). It is hypothesized that the addition of atezolizumab will yield a hazard ratio of 0.62 for PFS, for a sample size of 280 patients in the phase II portion of this study. The overall sample size for phase II/III will be 506, with the OS analysis designed to provide at least 85% power to detect a hazard ratio of 0.71 at a 1-sided significance level of 0.025. Secondary endpoints include objective response rates, local control, distant metastases free, and quality of life. This study includes a robust translational science component including blood and tissue based assays to further understand which patients may benefit most from immunotherapy. This study activated in May 2019 and is currently enrolling patients. NRG Oncology/Alliance LU005 is a randomized II/III trial testing the addition of atezolizumab to standard chemoradiation for LS-SCLC. The estimated date of study completion is May 2024. *Authors Higgins and Ross are co-first authors and contributed equally to this work.

Background Sarcoma mortality remains high despite adjuvant chemotherapy. Biomarker predictors of treatment response and outcome could improve treatment selection. Methods Tissue microarrays (TMAs) were created using pre- and posttreatment tumor from two prospective trials (MGH pilot and RTOG 9514) of neoadjuvant/adjuvant MAID chemotherapy and preoperative radiation. Biomarkers were measured using automated computerized imaging (AQUA or ACIS). Expression was correlated with disease-free survival (DFS), distant disease-free survival (DDFS), and overall survival (OS). Results Specimens from 60 patients included 23 pretreatment (PRE), 40 posttreatment (POST), and 12 matched pairs (MPs). In the MP set, CAIX, GLUT1, and PARP1 expression significantly decreased following neoadjuvant therapy, but p53 nuclear/cytoplasmic (N/C) ratio increased. In the PRE set, no biomarker expression was associated with DFS, DDFS, or OS. In the POST set, increased p53 N/C ratio was associated with a significantly decreased DFS and DDFS (HR 4.13, p=0.017; HR 4.16, p=0.016), while increased ERCC1 and XPF expression were associated with an improved DFS and DDFS. No POST biomarkers were associated with OS. Conclusions PRE biomarker expression did not predict survival outcomes. Expression pattern changes after neoadjuvant chemoradiation supports the concepts of tumor reoxygenation, altered HIF-1α signaling, and a p53 nuclear accumulation DNA damage response. Clinical Trial Registration NRG Oncology RTOG 9514 is registered with ClinicalTrials.gov. The ClinicalTrials.gov Identifier is NCT00002791.

343 Background: Biomarkers are needed to help select patients (pts) with muscle invading bladder cancer (MIBC) for bladder sparing chemotherapy and radiation treatment (CRT). Higher MRE11 expression has been identified as a potential RT response marker in MIBC. MRE11 protein is involved in the DNA double strand break repair mechanism. This analysis evaluates associations between MRE11 expression and outcome in pts from 6 NRG/RTOG bladder-sparing RT protocols. Methods: Archival tissue via TMA or unstained slides was used. Cases were stained with anti MRE11 antibody Rabbit mAb, clone EPR3471 (Epitomics at 1:1500 dilution). Slides were scanned on an Aperio FL instrument and analyzed via Automated Quantitative Image analysis (AQUA). MRE11 scores were determined within the nucleus and cytoplasm of urothelial cells and a ratio of nuclear to cytoplasmic (N/C) score calculated. A ratio was used to normalize scores and overcome pre-analytical variation. MRE11 N/C was analyzed by quartile cut points. Cumulative incidence was used to estimate disease-specific mortality (DSM; failure=bladder cancer death) and Fine-Gray models were used to evaluate associations between MRE11 and DSM. Cox models were used for overall survival (OS; death) and bladder-intact survival (BIS; cystectomy/death). Results: Out of 465 eligible pts, tissue was available and MRE11 N/C determined for 135. Analyzable pts were less likely to be white (p=0.0001) and more likely to be T2 (p=0.0003). Median MRE11 N/C was 2.41 (min-max: 0.69-6.03). Pts with MRE11 N/C ≤ 1.49 (lower quartile) were associated with significantly higher DSM (HR= 2, 95% CI: 1.1, 3.8, p=0.03). The 4-year DSM was 41% for pts with MER11 N/C ≤ 1.49 vs. 21% for pts with MER11 N/C was >1.49. MRE11 N/C was not associated with OS or BIS. Conclusions: AQUA analysis allows precise measurement of this marker in tissue samples. Low expression of MRE11 N/C (≤1.49) is associated with significantly higher DSM. This adds further evidence of MRE11 as a potential RT response biomarker for selection of pts most likely to respond to bladder-sparing CRT. Supported by NCI grants U10CA180868, U10CA180822, UG1CA189867,U24CA196067.

Expanding interest in and use of active surveillance for early state prostate cancer (PC) has increased need for prognostic biomarkers. Using a multi-institutional tissue microarray resource including over 1000 radical prostatectomy samples, we sought to correlate Ki67 expression captured by an automated image analysis system with clinicopathological features and validate its utility as a clinical grade test in predicting cancer-specific outcomes. After immunostaining, the Ki67 proliferation index (PI) of tumor areas of each core (three cancer cores/case) was analyzed using a nuclear quantification algorithm (Aperio). We assessed whether Ki67 PI was associated with clinicopathological factors and recurrence-free survival (RFS) including biochemical recurrence, metastasis or PC death (7-year median follow-up). In 1004 PCs (∼4000 tissue cores) Ki67 PI showed significantly higher inter-tumor (0.68) than intra-tumor variation (0.39). Ki67 PI was associated with stage (P<0.0001), seminal vesicle invasion (SVI, P=0.02), extracapsular extension (ECE, P<0.0001) and Gleason score (GS, P<0.0001). Ki67 PI as a continuous variable significantly correlated with recurrence-free, overall and disease-specific survival by multivariable Cox proportional hazard model (hazards ratio (HR)=1.04-1.1, P=0.02–0.0008). High Ki67 score (defined as ⩾5%) was significantly associated with worse RFS (HR=1.47, P=0.0007) and worse overall survival (HR=2.03, P=0.03). In localized PC treated by radical prostatectomy, higher Ki67 PI assessed using a clinical grade automated algorithm is strongly associated with a higher GS, stage, SVI and ECE and greater probability of recurrence.

Array comparative genomic hybridization was used to identify copy number alterations in clear cell renal cell carcinoma (ccRCC) patient tumors to identify associations with patient/clinical characteristics. Of 763 ccRCC patients, 412 (54%) provided frozen biopsies. Clones were analyzed for significant copy number differences, adjusting for multiple comparisons and covariates in multivariate analyses. Frequent alterations included losses on: 3p (92.2%), 14q (46.8%), 8p (38.1%), 4q (35.4%), 9p (32.3%), 9q (31.8%), 6q (30.8%), 3q (29.4%), 10q (25.7%), 13q (24.5%), 1p (23.5%) and gains on 5q (60.2%), 7q (39.6%), 7p (30.6%), 5p (26.5%), 20q (25.5%), 12q (24.8%), 12p (22.8%). Stage and grade were associated with 1p, 9p, 9q, 13q and 14q loss and 12q gain. Males had more alterations compared with females, independent of stage and grade. Significant differences in the number/types of alterations were observed by family cancer history, age at diagnosis and smoking status. Von Hippel-Lindau (VHL) gene inactivation was associated with 3p loss (P<E-05), and these cases had fewer alterations than wild-type cases. The fragile site flanking the FHIT locus (3p14.2) represented a unique breakpoint among VHL hypermethylated cases, compared with wild-type cases and those with sequence changes. This is the first study of its size to investigate copy number alterations among cases with extensive patient, clinical/risk factor information. Patients characterized by VHL wild-type gene status (vs sequence alterations) and male (vs female) cases had more copy number alterations regardless of diagnostic stage and grade, which could relate to poor prognosis.

2578 Background: CTL-associated antigen 4 (CTLA4) is a costimulatory molecule expressed on activated T cells that delivers an inhibitory signal to these T cells thereby serving as an immune checkpoint. Blocking CTLA4 with antibody treatment can induce tumor regression in cancer patients. While CTLA4 blockade can induce significant clinical responses in advanced prostate cancer patients, the antigen specific recognition that mediate these responses are unknown. Identification of these antigens may provide new insights into the immune recognition of prostate cancer including the development of future vaccines. Methods: We conducted a phase I clinical trial of anti-CTLA4 antibody (ipilimumab) combined with GM-CSF (sargramostim) in patients with castration resistant prostate cancer. To define the endogenous antigens against which immune responses are induced, we screened sera from clinical responders and nonresponders for treatment-induced IgG antibodies with arrays spotted with over 8,000 human proteins. We confirmed immune responses to these antigens by western blotting as well as detection of T cell immune responses to these antigens in the study subjects. We also assessed prostate cancer tissue for expression of these antigens. Results: We found that antibodies to a total of 117 antigens were increased in the clinical responders' sera after treatment, compared to a total of 31 antigens in the nonresponders. Some of the antigens were shared between some of the clinical responders. One of the shared antigens identified, Pak6, was found to be overexpressed at high frequency in prostate cancer, including biopsied tissue from a clinical responder who developed an immune response to this antigen. Conclusions: These results demonstrate that CTLA4 blockade can enhance tumor immune surveillance in vivo, enhancing immune responses to potentially common endogenous tumor antigens. No significant financial relationships to disclose.

S. M. Noworolski, R. F. Guo, G. D. Reed, K. Kuchinsky, K. Greene, P. Carroll, D. B. Vigneron, J. Kurhanewicz, and J. Simko Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, CA, United States, Graduate Group in Bioengineering, University of California, San Francisco & Berkeley, San Francisco & Berkeley, CA, United States, Electrical Engineering, University of California, Berkeley, Berkeley, CA, United States, Pathology, University of California, San Francisco, San Francisco, CA, United States, Urology, University of California, San Francisco, San Francisco, CA, United States

3063 Background: Granulocyte-macrophage colony-stimulating factor (GM-CSF) is a growth factor for antigen presenting cells that may enhance anti-tumor immune responses. Although this cytokine is being evaluated in immunotherapy trials for different malignancies, the impact of GM-CSF dose and schedule has not been directly addressed. Because systemic GM-CSF could potentiate anti-tumor immunity by enhancing presentation of endogenous tumor antigens, we examined whether increasing duration of GM-CSF treatment prior to radical prostatectomy could alter the composition of antigen presenting cells within the prostate tissue. Methods: A phase I trial in patients with localized prostate cancer was performed in subjects who were planned to undergo radical prostatectomy (RP). Two cohorts of 6 patients were sequentially accrued and treated with either 2 weeks or 3 weeks of daily GM-CSF (sargramostim, Bayer) 250 mg/m2/d SC. Patients were monitored for toxicity as well as for changes in circulating immune cells. RP tissues were assessed for immune cells by flow cytometry or immunohistochemistry (IHC) and digital slide scanning. In addition, IHC was also performed on RPs from untreated control Gleason grade- matched patients. Results: 12 patients were treated with neoadjuvant GM-CSF for either 2 weeks or 3 weeks. While significant leukocytosis was observed with treatment, no dose-limiting toxicities were encountered. As expected, circulating monocytes and dendritic cells (DC) were expanded following treatment. Compared to RPs from 11 untreated control patients, there was an increase in prostate-resident CD1a+ interstitial DC (Kruskal-Wallis test: p = 0.04). The increase was most pronounced in the 3-week treatment cohort. Moreover, tumor-resident DC expressed higher levels of the activation markers CD83 and 86 compared to DC residing in adjacent normal prostate and benign prostatic hypertrophy tissue. Conclusions: Systemically administered neoadjuvant GM-CSF can enhance the frequency of DC in prostate cancer. These effects appear to be dose dependent as they were most pronounced with 3 weeks of treatment. The composition of resident DC and monocytes also differs between malignant, benign, and hypertrophic prostate tissue. No significant financial relationships to disclose.

V. Zhang, M. Grinde, L. Tabatabai, J. Simko, M. Albers, D. Vigneron, and J. Kurhanewicz Radiology, University of California, San Francisco, San Francisco, CA, United States, Joint Bioengineering Program, University of California, Berkeley/San Francisco, San Francisco, CA, United States, Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway, St.Olav University Hospital, Trondheim, CA, Norway, Pathology, University of California, San Francisco, San Francisco, CA, United States

Prostate cancer is the second leading cause of cancer related deaths in US men, largely because of metastasis, which is ultimately fatal. A better understanding of metastasis biology will lead to improved prognostication and therapeutics. We previously reported 11q13.1 gain was independently predictive of recurrence after radical prostatectomy. Multiple endocrine neoplasia I (MEN1) maps to this region of copy number gain in aggressive prostate tumors and was shown to be the only gene at this locus at increased expression in prostate cancer. Here, we demonstrate an oncogenic role for MEN1 in prostate cancer using a variety of independent assays.

Introduction: The amount of nuclear antigen Ki-67 staining has been associated with cancer grade [1] and metastases [2]. Elevated levels of phospholipid metabolites have also been correlated with the presence of prostate cancer [3]. However, a direct correlation between the levels of individual phospholipid metabolites with cancer proliferation and Gleason grade has not been determined. This is of great clinical significance since most prostate cancers are indolent and will never metastasize, but there is currently no accurate way to identify these patients at diagnosis. HR-MAS twodimensional total correlation spectroscopy (TOCSY) can be used to quantify choline and ethanolamine phospholipid metabolites in intact human prostate tissues prior to pathology and immunohistochemistry of the same tissues [3]. The purpose of this study is to establish the relationship between phospholipid metabolite levels, Ki-67, and Gleason Grade. Methods: TOCSY’s were acquired from snap frozen tissue samples obtained at radical prostatectomy in order to resolve the choline (choline – Cho, phosphocholine – PC, glycerophosphocholine GPC) and ethanolamine-containing compounds (ethanolamine Eth, phosphoethanolamine PE, glycerolphosphethanolamine GPE), that overlap in 1-D HR-MAS tissue spectra. NMR spectra were acquired using a 11.7T (500MHz for H), 1°C, and a 2250 Hx spin rate using a Varian INOVA spectrometer equipped with a 4-mm gHX nanoprobe (Varian, Palo Alto, CA) and processed as previously described to generate absolute concentrations [3]. After HR-MAS, tissues were imbedded in OCT and 5 mm sections were obtained using a Leica CM1850 cryostat. For each sample, adjacent sections were stained with hematoxylin and eosin (H&E) and Ki-67 [1] and the % of the sample that was prostate cancer as well as the % of the cancer that stained positive for Ki-67 was determined by two experienced pathologists who were blinded to the clinical and metabolic findings. For the purpose of the abstract we divided cancers into low grade (≤ 3+4) and high grade (≥ 4+3).