External Validation of a Digital Pathology-based Multimodal Artificial Intelligence Architecture in the NRG/RTOG 9902 Phase 3 Trial.

BACKGROUND: Accurate risk stratification is critical to guide management decisions in localized prostate cancer (PCa). Previously, we had developed and validated a multimodal artificial intelligence (MMAI) model generated from digital histopathology and clinical features. Here, we externally validate this model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial. OBJECTIVE: To externally validate the MMAI model on men with high-risk or locally advanced PCa treated and followed as part of a phase 3 randomized control trial. DESIGN, SETTING, AND PARTICIPANTS: Our validation cohort included 318 localized high-risk PCa patients from NRG/RTOG 9902 with available histopathology (337 [85%] of the 397 patients enrolled into the trial had available slides, of which 19 [5.6%] failed due to poor image quality). OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Two previously locked prognostic MMAI models were validated for their intended endpoint: distant metastasis (DM) and PCa-specific mortality (PCSM). Individual clinical factors and the number of National Comprehensive Cancer Network (NCCN) high-risk features served as comparators. Subdistribution hazard ratio (sHR) was reported per standard deviation increase of the score with corresponding 95% confidence interval (CI) using Fine-Gray or Cox proportional hazards models. RESULTS AND LIMITATIONS: The DM and PCSM MMAI algorithms were significantly and independently associated with the risk of DM (sHR [95% CI] = 2.33 [1.60-3.38], p < 0.001) and PCSM, respectively (sHR [95% CI] = 3.54 [2.38-5.28], p < 0.001) when compared against other prognostic clinical factors and NCCN high-risk features. The lower 75% of patients by DM MMAI had estimated 5- and 10-yr DM rates of 4% and 7%, and the highest quartile had average 5- and 10-yr DM rates of 19% and 32%, respectively (p < 0.001). Similar results were observed for the PCSM MMAI algorithm. CONCLUSIONS: We externally validated the prognostic ability of MMAI models previously developed among men with localized high-risk disease. MMAI prognostic models further risk stratify beyond the clinical and pathological variables for DM and PCSM in a population of men already at a high risk for disease progression. This study provides evidence for consistent validation of our deep learning MMAI models to improve prognostication and enable more informed decision-making for patient care. PATIENT SUMMARY: This paper presents a novel approach using images from pathology slides along with clinical variables to validate artificial intelligence (computer-generated) prognostic models. When implemented, clinicians can offer a more personalized and tailored prognostic discussion for men with localized prostate cancer.

The Association of Radiation Dose With Overall Survival for Patients Treated With Prostate Stereotactic Body Radiation Therapy.

Introduction Stereotactic body radiation therapy (SBRT) for prostate adenocarcinoma (PCa) has demonstrated excellent biochemical recurrence-free survival, with studies showing improved BRFS with higher-dose SBRT. However, current studies have been underpowered to evaluate the relationship of SBRT dose to overall survival (OS). In this retrospective study using the National Cancer Database (NCDB), we hypothesize that, given the low alpha/beta ratio of PCa, a relatively small increase in the dose-per-fraction would be associated with improved survival outcomes for intermediate-risk PCa (IR-PCa) comparing 36.25 Gy/5 fx [biologically equivalent dose (BEDα/ß = 1.5 = 211.46 Gy vs. 35 Gy (BED1.5 = 198.33 Gy)]. Materials and methods We queried records from the NCDB from 2005 to 2015 for men receiving prostate SBRT for IR-PCa (n=2673). 82% were treated using either 35 Gy/5 fx or 36.25 Gy/5 fx. We compared OS in men receiving 35 Gy versus 36.25 Gy. Inverse probability of treatment weighting (IPTW) was used to adjust for covariable imbalances. Unweighted- and weighted-multivariable analysis (MVA) using Cox regression was used to compare OS hazard ratios, accounting for age, race, Charlson-Deyo comorbidity score, treatment facility type, prostate-specific antigen (PSA), clinical T-stage, Gleason Score, and use of androgen deprivation therapy (ADT). Kaplan-Meier analysis was performed. Results Seven hundred and eighty men (35%) were treated with 35 Gy/5 fx and 1434 men (65%) were treated with 36.25 Gy/5 fx (n=2214). Compared to 35 Gy, treatment with 36.25 Gy was associated with significantly improved OS (hazard ratio [HR]: 0.61 [95% CI: 0.43-0.89], P=0.009) on MVA. On Kaplan-Meier analysis, 36.25 Gy was associated with improved survival (p=0.034), with a five-year OS of 92% and 88%, respectively. Conclusions In a multi-institutional retrospective database of 2,214 IR patients treated with prostate SBRT, a prescription dose of 36.25 Gy/5 fx was associated with improved OS vs. 35 Gy/5 fx. Results are hypothesis-generating but do lend support to the current National Comprehensive Cancer Network (NCCN) guidelines that the minimum recommended dose for prostate SBRT is 36.25 Gy/5 fx.

Five-Year Patient-Reported Outcomes in NRG Oncology RTOG 0938, Evaluating Two Ultrahypofractionated Regimens for Prostate Cancer.

PURPOSE: There is considerable interest in very short (ultrahypofractionated) radiation therapy regimens to treat prostate cancer based on potential radiobiological advantages, patient convenience, and resource allocation benefits. Our objective is to demonstrate that detectable changes in health-related quality of life measured by the bowel and urinary domains of the Expanded Prostate Cancer Index Composite (EPIC-50) were not substantially worse than baseline scores. METHODS AND MATERIALS: NRG Oncology's RTOG 0938 is a nonblinded randomized phase 2 study of National Comprehensive Cancer Network low-risk prostate cancer in which each arm is compared with a historical control. Patients were randomized to 5 fractions (7.25 Gy in 2 week and a day [twice a week]) or 12 fractions (4.3Gy in 2.5 weeks [5 times a week]). Secondary objectives assessed patient-reported toxicity at 5 years using the EPIC. Chi-square tests were used to assess the proportion of patients with a deterioration from baseline of >5 points for bowel, >2 points for urinary, and >11 points for sexual score. RESULTS: The study enrolled 127 patients to 5 fractions (121 eligible) and 128 patients to 12 fractions (125 eligible). The median follow-up for all patients at the time of analysis was 5.38 years. The 5-year frequency for >5 point change in bowel score were 38.4% (P = .27) and 23.4% (P = 0.98) for 5 and 12 fractions, respectively. The 5-year frequencies for >2 point change in urinary score were 46.6% (P = .15) and 36.4% (P = .70) for 5 and 12 fractions, respectively. For 5 fractions, 49.3% (P = .007) of patients had a drop in 5-year EPIC-50 sexual score of ≥11 points; for 12 fractions, 54% (P < .001) of patients had a drop in 5-year EPIC-50 sexual score of ≥11 points. Disease-free survival at 5 years is 89.6% (95% CI: 84.0-95.2) in the 5-fraction arm and 92.3% (95% CI: 87.4-97.1) in the 12-fraction arm. There was no late grade 4 or 5 treatment-related urinary or bowel toxicity. CONCLUSIONS: This study confirms that, based on long-term changes in bowel and urinary domains and toxicity, the 5- and 12-fraction regimens are well tolerated. These ultrahypofractionated approaches need to be compared with current standard radiation therapy regimens.

Local Failure Events in Prostate Cancer Treated with Radiotherapy: A Pooled Analysis of 18 Randomized Trials from the Meta-analysis of Randomized Trials in Cancer of the Prostate Consortium (LEVIATHAN).

CONTEXT: The prognostic importance of local failure after definitive radiotherapy (RT) in National Comprehensive Cancer Network intermediate- and high-risk prostate cancer (PCa) patients remains unclear. OBJECTIVE: To evaluate the prognostic impact of local failure and the kinetics of distant metastasis following RT. EVIDENCE ACQUISITION: A pooled analysis was performed on individual patient data of 12 533 PCa (6288 high-risk and 6245 intermediate-risk) patients enrolled in 18 randomized trials (conducted between 1985 and 2015) within the Meta-analysis of Randomized Trials in Cancer of the Prostate Consortium. Multivariable Cox proportional hazard (PH) models were developed to evaluate the relationship between overall survival (OS), PCa-specific survival (PCSS), distant metastasis-free survival (DMFS), and local failure as a time-dependent covariate. Markov PH models were developed to evaluate the impact of specific transition states. EVIDENCE SYNTHESIS: The median follow-up was 11 yr. There were 795 (13%) local failure events and 1288 (21%) distant metastases for high-risk patients and 449 (7.2%) and 451 (7.2%) for intermediate-risk patients, respectively. For both groups, 81% of distant metastases developed from a clinically relapse-free state (cRF state). Local failure was significantly associated with OS (hazard ratio [HR] 1.17, 95% confidence interval [CI] 1.06-1.30), PCSS (HR 2.02, 95% CI 1.75-2.33), and DMFS (HR 1.94, 95% CI 1.75-2.15, p < 0.01 for all) in high-risk patients. Local failure was also significantly associated with DMFS (HR 1.57, 95% CI 1.36-1.81) but not with OS in intermediate-risk patients. Patients without local failure had a significantly lower HR of transitioning to a PCa-specific death state than those who had local failure (HR 0.32, 95% CI 0.21-0.50, p < 0.001). At later time points, more distant metastases emerged after a local failure event for both groups. CONCLUSIONS: Local failure is an independent prognosticator of OS, PCSS, and DMFS in high-risk and of DMFS in intermediate-risk PCa. Distant metastasis predominantly developed from the cRF state, underscoring the importance of addressing occult microscopic disease. However a "second wave" of distant metastases occurs subsequent to local failure events, and optimization of local control may reduce the risk of distant metastasis. PATIENT SUMMARY: Among men receiving definitive radiation therapy for high- and intermediate-risk prostate cancer, about 10% experience local recurrence, and they are at significantly increased risks of further disease progression. About 80% of patients who develop distant metastasis do not have a detectable local recurrence preceding it.

Patient Reported Outcomes in NRG Oncology RTOG 0938, Evaluating Two Ultrahypofractionated Regimens for Prostate Cancer.

PURPOSE: There is considerable interest in very short (ultrahypofractionated) radiation therapy regimens to treat prostate cancer based on potential radiobiological advantages, patient convenience, and resource allocation benefits. Our objective is to demonstrate that detectable changes in health-related quality of life measured by the bowel and urinary domains of the Expanded Prostate Cancer Index Composite (EPIC-50) were not substantially worse than baseline scores. METHODS AND MATERIALS: NRG Oncology's RTOG 0938 is a nonblinded randomized phase 2 study of National Comprehensive Cancer Network low-risk prostate cancer in which each arm is compared with a historical control. Patients were randomized to 5 fractions (7.25 Gy in 2 weeks) or 12 fractions (4.3 Gy in 2.5 weeks). The co-primary endpoints were the proportion of patients with a change in EPIC-50 bowel score at 1 year (baseline to 1 year) >5 points and in EPIC-50 urinary score >2 points tested with a 1-sample binomial test. RESULTS: The study enrolled 127 patients to 5 fractions (121 analyzed) and 128 patients to 12 fractions (125 analyzed). Median follow-up for all patients at the time of analysis was 3.8 years. The 1-year frequency for >5 point change in bowel score were 29.8% (P < .001) and 28.4% (P < .001) for 5 and 12 fractions, respectively. The 1-year frequencies for >2 point change in urinary score were 45.7% (P < .001) and 42.2% (P < .001) for 5 and 12 fractions, respectively. For 5 fractions, 32.9% of patients had a drop in 1-year EPIC-50 sexual score of ≥11 points (P = .34); for 12 fractions, 30.9% of patients had a drop in 1-year EPIC-50 sexual score of ≥ 11 points (P = .20). Disease-free survival at 2 years is 99.2% (95% confidence interval: 97.5-100) in the 5-fraction arm and 97.5% (95% confidence interval: 94.6-100) in the 12-fraction arm. There was no late grade 4 or 5 treatment-related urinary or bowel toxicity. CONCLUSIONS: This study confirms that, based on changes in bowel and urinary domains and toxicity (acute and late), the 5- and 12-fraction regimens are well tolerated. These ultrahypofractionated approaches need to be compared with current standard radiation therapy regimens.