Comparative Evaluation of LLMs in Clinical Oncology.

BACKGROUND: As artificial intelligence (AI) tools become widely accessible, more patients and medical professionals will turn to them for medical information. Large language models (LLMs), a subset of AI, excel in natural language processing tasks and hold considerable promise for clinical use. Fields such as oncology, in which clinical decisions are highly dependent on a continuous influx of new clinical trial data and evolving guidelines, stand to gain immensely from such advancements. It is therefore of critical importance to benchmark these models and describe their performance characteristics to guide their safe application to clinical oncology. Accordingly, the primary objectives of this work were to conduct comprehensive evaluations of LLMs in the field of oncology and to identify and characterize strategies that medical professionals can use to bolster their confidence in a model's response. METHODS: This study tested five publicly available LLMs (LLaMA 1, PaLM 2, Claude-v1, generative pretrained transformer 3.5 [GPT-3.5], and GPT-4) on a comprehensive battery of 2044 oncology questions, including topics from medical oncology, surgical oncology, radiation oncology, medical statistics, medical physics, and cancer biology. Model prompts were presented independently of each other, and each prompt was repeated three times to assess output consistency. For each response, models were instructed to provide a self-appraised confidence score (from 1 to 4). Model performance was also evaluated against a novel validation set comprising 50 oncology questions curated to eliminate any risk of overlap with the data used to train the LLMs. RESULTS: There was significant heterogeneity in performance between models (analysis of variance, P<0.001). Relative to a human benchmark (2013 and 2014 examination results), GPT-4 was the only model to perform above the 50th percentile. Overall, model performance varied as a function of subject area across all models, with worse performance observed in clinical oncology subcategories compared with foundational topics (medical statistics, medical physics, and cancer biology). Within the clinical oncology subdomain, worse performance was observed in female-predominant malignancies. A combination of model selection, prompt repetition, and confidence self-appraisal allowed for the identification of high-performing subgroups of questions with observed accuracies of 81.7 and 81.1% in the Claude-v1 and GPT-4 models, respectively. Evaluation of the novel validation question set produced similar trends in model performance while also highlighting improved performance in newer, centrally hosted models (GPT-4 Turbo and Gemini 1.0 Ultra) and local models (Mixtral 8×7B and LLaMA 2). CONCLUSIONS: Of the models tested on a standardized set of oncology questions, GPT-4 was observed to have the highest performance. Although this performance is impressive, all LLMs continue to have clinically significant error rates, including examples of overconfidence and consistent inaccuracies. Given the enthusiasm to integrate these new implementations of AI into clinical practice, continued standardized evaluations of the strengths and limitations of these products will be critical to guide both patients and medical professionals. (Funded by the National Institutes of Health Clinical Center for Research and the Intramural Research Program of the National Institutes of Health; Z99 CA999999.).

Target Volume Optimization for Localized Prostate Cancer.

PURPOSE: To provide a comprehensive review of the means by which to optimize target volume definition for the purposes of treatment planning for patients with intact prostate cancer with a specific emphasis on focal boost volume definition. METHODS: Here we conduct a narrative review of the available literature summarizing the current state of knowledge on optimizing target volume definition for the treatment of localized prostate cancer. RESULTS: Historically, the treatment of prostate cancer included a uniform prescription dose administered to the entire prostate with or without coverage of all or part of the seminal vesicles. The development of prostate magnetic resonance imaging (MRI) and positron emission tomography (PET) using prostate-specific radiotracers has ushered in an era in which radiation oncologists are able to localize and focally dose-escalate high-risk volumes in the prostate gland. Recent phase 3 data has demonstrated that incorporating focal dose escalation to high-risk subvolumes of the prostate improves biochemical control without significantly increasing toxicity. Still, several fundamental questions remain regarding the optimal target volume definition and prescription strategy to implement this technique. Given the remaining uncertainty, a knowledge of the pathological correlates of radiographic findings and the anatomic patterns of tumor spread may help inform clinical judgement for the definition of clinical target volumes. CONCLUSION: Advanced imaging has the ability to improve outcomes for patients with prostate cancer in multiple ways, including by enabling focal dose escalation to high-risk subvolumes. However, many questions remain regarding the optimal target volume definition and prescription strategy to implement this practice, and key knowledge gaps remain. A detailed understanding of the pathological correlates of radiographic findings and the patterns of local tumor spread may help inform clinical judgement for target volume definition given the current state of uncertainty.

Dosimetric evaluation and treatment planning considerations for GammaTile permanent brain implants - a pilot, institutional experience.

PURPOSE: GammaTile® (GT) is a brachytherapy platform that received Federal Drug Administration (FDA) approval as brain tumor therapy in late 2018. Here, we reviewed our institutional experience with GT as treatment for recurrent glioblastomas and characterized dosimetric parameter and associated clinical outcome. METHODS AND MATERIALS: A total of 20 consecutive patients with 21 (n = 21) diagnosis of recurrent glioblastoma underwent resection followed by intraoperative GT implant between 01/2019 and 12/2020. Data on gross tumor volume (GTV), number of GT units implanted, dose coverage for the high-risk clinical target volume (HR-CTV), measured by D90 or dose received by 90% of the HR-CTV, dose to organs at risk, and six months local control were collected. RESULTS: The median D90 to HR-CTV was 56.0 Gy (31.7-98.7 Gy). The brainstem, optic chiasm, ipsilateral optic nerve, and ipsilateral hippocampus median Dmax were 11.2, 5.4, 6.4, and 10.0 Gy, respectively. None of the patients in this study cohort suffered from radiation necrosis or adverse events attributable to the GT. Correlation was found between pre-op GTV, the volume of the resection cavity, and the number of GT units implanted. Of the resection cavities, 7/21 (33%) of the cavity experienced shrinkage, 3/21 (14%) remained stable, and 11/21 (52%) of the cavities expanded on the 3-months post-resection/GT implant MRIs. D90 to HR-CTV was found to be associated with local recurrence at 6-month post GT implant, suggesting a dose response relationship (p = 0.026). The median local recurrence-free survival was 366.5 days (64-1,098 days), and a trend towards improved local recurrence-free survival was seen in patients with D90 to HR-CTV ≥ 56 Gy (p = 0.048). CONCLUSIONS: Our pilot, institutional experience provides clinical outcome, dosimetric considerations, and offer technical guidance in the clinical implementation of GT brachytherapy.

Automated prostate gland segmentation in challenging clinical cases: comparison of three artificial intelligence methods.

OBJECTIVE: Automated methods for prostate segmentation on MRI are typically developed under ideal scanning and anatomical conditions. This study evaluates three different prostate segmentation AI algorithms in a challenging population of patients with prior treatments, variable anatomic characteristics, complex clinical history, or atypical MRI acquisition parameters. MATERIALS AND METHODS: A single institution retrospective database was queried for the following conditions at prostate MRI: prior prostate-specific oncologic treatment, transurethral resection of the prostate (TURP), abdominal perineal resection (APR), hip prosthesis (HP), diversity of prostate volumes (large ≥ 150 cc, small ≤ 25 cc), whole gland tumor burden, magnet strength, noted poor quality, and various scanners (outside/vendors). Final inclusion criteria required availability of axial T2-weighted (T2W) sequence and corresponding prostate organ segmentation from an expert radiologist. Three previously developed algorithms were evaluated: (1) deep learning (DL)-based model, (2) commercially available shape-based model, and (3) federated DL-based model. Dice Similarity Coefficient (DSC) was calculated compared to expert. DSC by model and scan factors were evaluated with Wilcox signed-rank test and linear mixed effects (LMER) model. RESULTS: 683 scans (651 patients) met inclusion criteria (mean prostate volume 60.1 cc [9.05-329 cc]). Overall DSC scores for models 1, 2, and 3 were 0.916 (0.707-0.971), 0.873 (0-0.997), and 0.894 (0.025-0.961), respectively, with DL-based models demonstrating significantly higher performance (p < 0.01). In sub-group analysis by factors, Model 1 outperformed Model 2 (all p < 0.05) and Model 3 (all p < 0.001). Performance of all models was negatively impacted by prostate volume and poor signal quality (p < 0.01). Shape-based factors influenced DL models (p < 0.001) while signal factors influenced all (p < 0.001). CONCLUSION: Factors affecting anatomical and signal conditions of the prostate gland can adversely impact both DL and non-deep learning-based segmentation models.

A Phase 1 Trial of Salvage Stereotactic Body Radiation Therapy for Radiorecurrent Prostate Cancer After Brachytherapy.

PURPOSE: NCT03253744 is a phase 1 trial with the primary objective to identify the maximum tolerated dose (MTD) of salvage stereotactic body radiation therapy (SBRT) in patients with local prostate cancer recurrence after brachytherapy. Additional objectives included biochemical control and imaging response. METHODS AND MATERIALS: This trial was initially designed to test 3 therapeutic dose levels (DLs): 40 Gy (DL1), 42.5 Gy (DL2), and 45 Gy (DL3) in 5 fractions. Intensity modulation was used to deliver the prescription dose to the magnetic resonance imaging and prostate-specific membrane antigen-based positron emission tomography imaging-defined gross tumor volume while simultaneously delivering 30 Gy to an elective volume defined by the prostate gland. This phase 1 trial followed a 3+3 design with a 3-patient expansion at the MTD. Toxicities were scored until trial completion at 2 years post-SBRT using Common Terminology Criteria for Adverse Events version 5.0. Escalation was halted if 2 dose limiting toxicities occurred, defined as any persistent (>4 days) grade 3 toxicity occurring within the first 3 weeks after SBRT or any grade ≥3 genitourinary (GU) or grade 4 gastrointestinal toxicity thereafter. RESULTS: Between August 2018 and January 2023, 9 patients underwent salvage SBRT and were observed for a median of 22 months (Q1-Q3, 20-43 months). No grade 3 to 5 adverse events related to study treatment were observed; thus, no dose limiting toxicities occurred during the observation period. Escalation was halted by amendment given excellent biochemical control in DL1 and DL2 in the setting of a high incidence of clinically significant late grade 2 GU toxicity. Therefore, the MTD was considered 42.5 Gy in 5 fractions (DL2). One- and 2-year biochemical progression-free survival were 100% and 86%, representing a single patient in the trial cohort with biochemical failure (prostate-specific antigen [PSA] nadir + 2.0) at 20 months posttreatment. CONCLUSIONS: The MTD of salvage SBRT for the treatment of intraprostatic radiorecurrence after brachytherapy was 42.5 Gy in 5 fractions producing an 86% 2-year biochemical progression-free survival rate, with 1 poststudy failure at 20 months. The most frequent clinically significant toxicity was late grade 2 GU toxicity.

A platform-independent AI tumor lineage and site (ATLAS) classifier.

Histopathologic diagnosis and classification of cancer plays a critical role in guiding treatment. Advances in next-generation sequencing have ushered in new complementary molecular frameworks. However, existing approaches do not independently assess both site-of-origin (e.g. prostate) and lineage (e.g. adenocarcinoma) and have minimal validation in metastatic disease, where classification is more difficult. Utilizing gradient-boosted machine learning, we developed ATLAS, a pair of separate AI Tumor Lineage and Site-of-origin models from RNA expression data on 8249 tumor samples. We assessed performance independently in 10,376 total tumor samples, including 1490 metastatic samples, achieving an accuracy of 91.4% for cancer site-of-origin and 97.1% for cancer lineage. High confidence predictions (encompassing the majority of cases) were accurate 98-99% of the time in both localized and remarkably even in metastatic samples. We also identified emergent properties of our lineage scores for tumor types on which the model was never trained (zero-shot learning). Adenocarcinoma/sarcoma lineage scores differentiated epithelioid from biphasic/sarcomatoid mesothelioma. Also, predicted lineage de-differentiation identified neuroendocrine/small cell tumors and was associated with poor outcomes across tumor types. Our platform-independent single-sample approach can be easily translated to existing RNA-seq platforms. ATLAS can complement and guide traditional histopathologic assessment in challenging situations and tumors of unknown primary.

Predicting 18F-DCFPyL-PET/CT Scan Positivity in Prostate Cancer Patients with Biochemical Recurrence.

RATIONALE AND OBJECTIVES: To analyze variables that can predict the positivity of 18F-DCFPyL- positron emission tomography/computed tomography (PET/CT) and extent of disease in patients with biochemically recurrent (BCR) prostate cancer after primary local therapy with either radical prostatectomy or radiation therapy. MATERIALS AND METHODS: This is a retrospective analysis of a prospective single institutional review board-approved study. We included 199 patients with biochemical recurrence and negative conventional imaging after primary local therapies (radical prostatectomy n = 127, radiation therapy n = 72). All patients underwent 18F-DCFPyL-PET/CT. Univariate and multivariate logistic regression analyses were used to determine predictors of a positive scan for both cohort of patients. Regression-based coefficients were used to develop nomograms predicting scan positivity and extra-pelvic disease. Decision curve analysis (DCA) was implemented to quantify nomogram's clinical benefit. RESULTS: Of the 127 (63%) post-radical prostatectomy patients, 91 patients had positive scans - 61 of those with intrapelvic lesions and 30 with extra-pelvic lesions (i.e., retroperitoneal or distant nodes and/or bone/organ lesions). Of the 72 post-radiation therapy patients, 65 patients had positive scans - 39 of them had intrapelvic lesions and 26 extra-pelvic lesions. In the radical prostatectomy cohort, multivariate regression analysis revealed original International Society of Urological Pathology category, prostate-specific antigen (PSA), prostate-specific antigen doubling time (PSAdt), and time from BCR (mo) to scan were predictors for scan positivity and presence of extra-pelvic disease, with an area under the curve of 80% and 78%, respectively. Positive versus negative tumor margin after radical prostatectomy was not related to scan positivity or to the presence of positive extra-pelvic foci. In the radiation therapy cohort, multivariate regression analysis revealed that PSA, PSAdt, and time to BCR (mo) were predictors of extra-pelvic disease, with area under the curve of 82%. Because only seven patients in the radiation therapy cohort had negative scans, a prediction model for scan positivity could not be analyzed and only the presence of extra-pelvic disease was evaluated. CONCLUSION: PSA and PSAdt are consistently significant predictors of 18F-DCFPyL PET/CT positivity and extra-pelvic disease in BCR prostate cancer patients. Stratifying the patient population into primary local treatment group enables the use of other variables as predictors, such as time since BCR. This nomogram may guide selection of the most suitable candidates for 18F-DCFPyL-PET/CT imaging.

PI-RADS Version 2.0 Versus Version 2.1: Comparison of Prostate Cancer Gleason Grade Upgrade and Downgrade Rates From MRI-Targeted Biopsy to Radical Prostatectomy.

BACKGROUND. Precise risk stratification through MRI/ultrasound (US) fusion-guided targeted biopsy (TBx) can guide optimal prostate cancer (PCa) management. OBJECTIVE. The purpose of this study was to compare PI-RADS version 2.0 (v2.0) and PI-RADS version 2.1 (v2.1) in terms of the rates of International Society of Urological Pathology (ISUP) grade group (GG) upgrade and downgrade from TBx to radical prostatectomy (RP). METHODS. This study entailed a retrospective post hoc analysis of patients who underwent 3-T prostate MRI at a single institution from May 2015 to March 2023 as part of three prospective clinical trials. Trial participants who underwent MRI followed by MRI/US fusion-guided TBx and RP within a 1-year interval were identified. A single genitourinary radiologist performed clinical interpretations of the MRI examinations using PI-RADS v2.0 from May 2015 to March 2019 and PI-RADS v2.1 from April 2019 to March 2023. Upgrade and downgrade rates from TBx to RP were compared using chi-square tests. Clinically significant cancer was defined as ISUP GG2 or greater. RESULTS. The final analysis included 308 patients (median age, 65 years; median PSA density, 0.16 ng/mL2). The v2.0 group (n = 177) and v2.1 group (n = 131) showed no significant difference in terms of upgrade rate (29% vs 22%, respectively; p = .15), downgrade rate (19% vs 21%, p = .76), clinically significant upgrade rate (14% vs 10%, p = .27), or clinically significant downgrade rate (1% vs 1%, p > .99). The upgrade rate and downgrade rate were also not significantly different between the v2.0 and v2.1 groups when stratifying by index lesion PI-RADS category or index lesion zone, as well as when assessed only in patients without a prior PCa diagnosis (all p > .01). Among patients with GG2 or GG3 at RP (n = 121 for v2.0; n = 103 for v2.1), the concordance rate between TBx and RP was not significantly different between the v2.0 and v2.1 groups (53% vs 57%, p = .51). CONCLUSION. Upgrade and downgrade rates from TBx to RP were not significantly different between patients whose MRI examinations were clinically interpreted using v2.0 or v2.1. CLINICAL IMPACT. Implementation of the most recent PI-RADS update did not improve the incongruence in PCa grade assessment between TBx and surgery.

Evaluation of a Deep Learning-based Algorithm for Post-Radiotherapy Prostate Cancer Local Recurrence Detection Using Biparametric MRI.

OBJECTIVE: To evaluate a biparametric MRI (bpMRI)-based artificial intelligence (AI) model for the detection of local prostate cancer (PCa) recurrence in patients with radiotherapy history. MATERIALS AND METHODS: This study included post-radiotherapy patients undergoing multiparametric MRI and subsequent MRI/US fusion-guided and/or systematic biopsy. Histopathology results were used as ground truth. The recurrent cancer detection sensitivity of a bpMRI-based AI model, which was developed on a large dataset to primarily identify lesions in treatment-naïve patients, was compared to a prospective radiologist assessment using the Wald test. Subanalysis was conducted on patients stratified by the treatment modality (external beam radiation treatment [EBRT] and brachytherapy) and the prostate volume quartiles. RESULTS: Of the 62 patients included (median age = 70 years; median PSA = 3.51 ng/ml; median prostate volume = 27.55 ml), 56 recurrent PCa foci were identified within 46 patients. The AI model detected 40 lesions in 35 patients. The AI model performance was lower than the prospective radiology interpretation (Rad) on a patient-(AI: 76.1% vs. Rad: 91.3%, p = 0.02) and lesion-level (AI: 71.4% vs. Rad: 87.5%, p = 0.01). The mean number of false positives per patient was 0.35 (range: 0-2). The AI model performance was higher in EBRT group both on patient-level (EBRT: 81.5% [22/27] vs. brachytherapy: 68.4% [13/19]) and lesion-level (EBRT: 79.4% [27/34] vs. brachytherapy: 59.1% [13/22]). In patients with gland volumes >34 ml (n = 25), detection sensitivities were 100% (11/11) and 94.1% (16/17) on patient- and lesion-level, respectively. CONCLUSION: The reported bpMRI-based AI model detected the majority of locally recurrent prostate cancer after radiotherapy. Further testing including external validation of this model is warranted prior to clinical implementation.

A Phase 1 Trial of Focal Salvage Stereotactic Body Radiation Therapy for Radiorecurrent Prostate Cancer.

PURPOSE: NCT03253744 was a phase 1 trial to identify the maximum tolerated dose (MTD) of image-guided, focal, salvage stereotactic body radiation therapy (SBRT) for patients with locally radiorecurrent prostate cancer. Additional objectives included biochemical control and imaging response. METHODS AND MATERIALS: The trial design included 3 dose levels (DLs): 40 Gy (DL1), 42.5 Gy (DL2), and 45 Gy (DL3) in 5 fractions delivered ≥48 hours apart. The prescription dose was delivered to the magnetic resonance- and prostate-specific membrane antigen imaging-defined tumor volume. Dose escalation followed a 3+3 design with a 3-patient expansion at the MTD. Toxicities were scored until 2 years after completion of SBRT using Common Terminology Criteria for Adverse Events, version 5.0, criteria. Escalation was halted if 2 dose-limiting toxicities occurred, defined as any persistent (>4 days) grade 3 toxicity occurring within the first 3 weeks after SBRT and any grade 3 genitourinary (GU) or grade 4 gastrointestinal (GI) toxicity thereafter. RESULTS: Between August 2018 and May 2022, 8 patients underwent salvage focal SBRT, with a median follow-up of 35 months. No dose-limiting toxic effects were observed on DL1. Two patients were enrolled in DL2 and experienced grade 3 GU toxicities, prompting de-escalation and expansion (n = 6) at the MTD (DL1). The most common toxicities observed were grade ≥2 GU toxicities, with only a single grade 2 GI toxicity and no grade ≥3 GI toxicities. One patient experienced biochemical failure (prostate-specific antigen nadir + 2.0) at 33 months. CONCLUSIONS: The MTD for focal salvage SBRT for isolated intraprostatic radiorecurrence was 40 Gy in 5 fractions, producing a 100% 24-month biochemical progression free survival, with 1 poststudy failure at 33 months. The most frequent clinically significant toxicity was late grade ≥2 GU toxicity.

Safety of Radiation Therapy in Patients With Prostate Cancer and Inflammatory Bowel Disease: A Systematic Review.

PURPOSE: Inflammatory bowel disease (IBD) has historically been considered a relative contraindication for pelvic radiation therapy (RT). To date, no systematic review has summarized the toxicity profile of RT for patients with prostate cancer and comorbid IBD. METHODS AND MATERIALS: A PRISMA-guided systematic search was conducted on PubMed/Embase for original investigations that reported gastrointestinal (GI; rectal/bowel) toxicity in patients with IBD undergoing RT for prostate cancer. The substantial heterogeneity in patient population, follow-up, and toxicity reporting practices precluded a formal meta-analysis; however, a summary of the individual study-level data and crude pooled rates was described. RESULTS: Twelve retrospective studies with 194 patients were included: 5 examined predominantly low-dose-rate brachytherapy (BT) monotherapy, 1 predominantly high-dose-rate BT monotherapy, 3 mixed external beam RT (3-dimensional conformal or intensity modulated RT [IMRT]) + low-dose-rate BT, 1 IMRT + high-dose-rate BT, and 2 stereotactic RT. Among these studies, patients with active IBD, patients receiving pelvic RT, and patients with prior abdominopelvic surgery were underrepresented. In all but 1 publication, the rate of late grade 3+ GI toxicities was <5%. The crude pooled rate of acute and late grade 2+ GI events was 15.3% (n = 27/177 evaluable patients; range, 0%-100%) and 11.3% (n = 20/177 evaluable patients; range, 0%-38.5%), respectively. Crude rates of acute and late grade 3+ GI events were 3.4% (6 cases; range, 0%-23%) and 2.3% (4 cases; range, 0%-15%). CONCLUSIONS: Prostate RT in patients with comorbid IBD appears to be associated with low rates of grade 3+ GI toxicity; however, patients must be counseled regarding the possibility for lower-grade toxicities. These data cannot be generalized to the underrepresented subpopulations mentioned above, and individualize decision-making is recommended for those high-risk cases. Several strategies should be considered to minimize the probability of toxicity in this susceptible population, including careful patient selection, minimizing elective (nodal) treatment volumes, using rectal sparing techniques, and employing contemporary RT advancements to minimize exposure to GI organs at risk (eg, IMRT, magnetic resonance imaging-based target delineation, and high-quality daily image guidance).

A Phase 1 Trial of Highly Conformal, Hypofractionated Postprostatectomy Radiation Therapy.

Purpose: This phase 1 trial aimed to identify the maximally tolerated hypofractionated dose schedule for postoperative radiation therapy (PORT) after radical prostatectomy. Secondary objectives included biochemical control and quality of life (QoL) measures. Methods and Materials: Patients were treated on 1 of 3 dose levels (DLs): 56.4 Gy in 20 fractions (DL1), 51.2 Gy in 15 fractions (DL2), and 44.2 Gy in 10 fractions (DL3). Treatment was delivered to the prostate bed without pelvic nodal irradiation. Dose escalation followed a standard 3 + 3 design with an expansion for 6 additional patients at the maximally tolerated hypofractionated dose schedule. Acute dose-limiting toxicity (DLT) was defined as grade 3 toxicity lasting >4 days within 21 days of PORT completion; late DLT was defined as grade 4 gastrointestinal (GI) or genitourinary (GU) toxicity. Results: Between January 2018 and August 2019, 15 patients underwent radiation treatment: 3 on DL1, 3 on DL2, and 9 on DL3. The median follow-up was 24 months. There were no DLTs, and the maximally tolerated hypofractionated dose schedule was identified as DL3. Two of the 15 patients (13.3%) experienced biochemical failure (prostate-specific antigen >0.1). Ten of 15 patients (67%) had grade 2+ acute toxicities, consisting of transient GI toxicities. Three patients experienced late grade 2+ GI toxicity, and 5 patients experienced late grade 2+ GU toxicity. Late grade 3 GU toxicity occurred in 2 patients. There were no grade 4+ acute or late toxicities. There were no significant differences in GI measures of QoL, however, there was an increase in GU symptoms and corresponding decrease in GU QoL between 12 and 24 months. Conclusions: The maximum tolerated hypofractionated dose schedule for hypofractionated PORT to the prostate bed was determined to be 44.2 Gy in 10 daily fractions. The most frequent clinically significant toxicities were late grade 2+ GU toxicities, which corresponded to a worsening of late GU QoL.

Detection of failure patterns using advanced imaging in patients with biochemical recurrence following low-dose-rate brachytherapy for prostate cancer.

PURPOSE/OBJECTIVE(S): This study describes the pattern of failure in patients with biochemical (BCR) recurrence after low-dose-rate (LDR) brachytherapy as a component of definitive treatment for prostate cancer. METHODS: Patients with BCR after LDR brachytherapy ± external beam radiation therapy (EBRT) were enrolled on prospective IRB approved advanced imaging protocols. Patients underwent 3T multiparametric MRI (mpMRI); a subset underwent prostate specific membrane antigen (PSMA)-based PET/CT. Pathologic confirmation was obtained unless contraindicated. RESULTS: Between January 2011 and April 2021, 51 patients with BCR after brachytherapy (n = 36) or brachytherapy + EBRT (n = 15) underwent mpMRI and were included in this analysis. Of 38 patients with available dosimetry, only two had D90<90%. The prostate and seminal vesicles were a site of failure in 66.7% (n = 34) and 39.2% (n = 20), respectively. PET/CT (n = 32 patients) more often identified lesions pelvic lymph nodes (50%; n = 16) and distant metastases (18.8%; n = 6), than mpMRI. Isolated nodal disease (9.8%; n = 5) and distant metastases (n = 1) without local recurrence were uncommon. Recurrence within the prostate was located in the transition zone in 48.5%, central or midline in 45.5%, and anterior in 36.4% of patients. CONCLUSION: In this cohort of patients with BCR after LDR brachytherapy ± EBRT, the predominant recurrence pattern was local (prostate ± seminal vesicles) with frequent occurrence in the anterior prostate and transition zone. mpMRI and PSMA PET/CT provided complementary information to localize sites of recurrence, with PSMA PET/CT often confirming mpMRI findings and identifying occult nodal or distant metastases.

Development of a 3D CNN-based AI Model for Automated Segmentation of the Prostatic Urethra.

RATIONALE AND OBJECTIVE: The combined use of prostate cancer radiotherapy and MRI planning is increasingly being used in the treatment of clinically significant prostate cancers. The radiotherapy dosage quantity is limited by toxicity in organs with de-novo genitourinary toxicity occurrence remaining unperturbed. Estimation of the urethral radiation dose via anatomical contouring may improve our understanding of genitourinary toxicity and its related symptoms. Yet, urethral delineation remains an expert-dependent and time-consuming procedure. In this study, we aim to develop a fully automated segmentation tool for the prostatic urethra. MATERIALS AND METHODS: This study incorporated 939 patients' T2-weighted MRI scans (train/validation/test/excluded: 657/141/140/1 patients), including in-house and public PROSTATE-x datasets, and their corresponding ground truth urethral contours from an expert genitourinary radiologist. The AI model was developed using MONAI framework and was based on a 3D-UNet. AI model performance was determined by Dice score (volume-based) and the Centerline Distance (CLD) between the prediction and ground truth centers (slice-based). All predictions were compared to ground truth in a systematic failure analysis to elucidate the model's strengths and weaknesses. The Wilcoxon-rank sum test was used for pair-wise comparison of group differences. RESULTS: The overall organ-adjusted Dice score for this model was 0.61 and overall CLD was 2.56 mm. When comparing prostates with symmetrical (n = 117) and asymmetrical (n = 23) benign prostate hyperplasia (BPH), the AI model performed better on symmetrical prostates compared to asymmetrical in both Dice score (0.64 vs. 0.51 respectively, p < 0.05) and mean CLD (2.3 mm vs. 3.8 mm respectively, p < 0.05). When calculating location-specific performance, the performance was highest at the apex and lowest at the base location of the prostate for Dice and CLD. Dice location dependence: symmetrical (Apex, Mid, Base: 0.69 vs. 0.67 vs. 0.54 respectively, p < 0.05) and asymmetrical (Apex, Mid, Base: 0.68 vs. 0.52 vs. 0.39 respectively, p < 0.05). CLD location dependence: symmetrical (Apex, Mid, Base: 1.43 mm vs. 2.15 mm vs. 3.28 mm, p < 0.05) and asymmetrical (Apex, Mid, Base: 1.83 mm vs. 3.1 mm vs. 6.24 mm, p < 0.05). CONCLUSION: We developed a fully automated prostatic urethra segmentation AI tool yielding its best performance in prostate glands with symmetric BPH features. This system can potentially be used to assist treatment planning in patients who can undergo whole gland radiation therapy or ablative focal therapy.

Tissue-Based Biomarkers for the Risk Stratification of Men With Clinically Localized Prostate Cancer.

Risk stratification of men with clinically localized prostate cancer has historically relied on basic clinicopathologic parameters such as prostate specific antigen level, grade group, and clinical stage. However, prostate cancer often behaves in ways that cannot be accurately predicted by these parameters. Thus, recent efforts have focused on developing tissue-based genomic tests that provide greater insights into the risk of a given patient's disease. Multiple tests are now commercially available and provide additional prognostic information at various stages of the care pathway for prostate cancer. Indeed, early evidence suggests that these assays may have a significant impact on patient and physician decision-making. However, the impact of these tests on oncologic outcomes remains less clear. In this review, we highlight recent advances in the use of tissue-based biomarkers in the treatment of prostate cancer and identify the existing evidence supporting their clinical use.

Cutaneous T-cell lymphoma is associated with increased risk of lymphoma, melanoma, lung cancer, and bladder cancer.

BACKGROUND: Patients with cutaneous T-cell lymphoma (CTCL) are at a higher risk of developing second malignancies. However, rates of incidence vary significantly across studies. METHODS: A systematic review and meta-analysis of articles published between 1950 and 2019 was performed to evaluate the risk of second malignancies in patients with CTCL. RESULTS: We identified 10 eligible studies, including 12 patient cohorts, with 5.9% to 16.8% of patients developing second malignancies. All studies showed a male predominance for patients developing second malignancies. The mean age across the studies ranged from 44.6 to 68.0 years. The time between the diagnosis of CTCL and second malignancy ranged from 2.1 to 5.4 years (mean, 3.29 y; 95% confidence interval [CI], 2.69-5.15). Meta-analysis showed a standardized incidence ratio of 2.18 (95% CI, 1.43-2.93) for all malignancies. The standardized incidence ratios were 15.25 (95% CI, 7.70-22.79) for Hodgkin lymphoma, 4.96 (95% CI, 3.58-6.33) for non-Hodgkin lymphoma, 1.69 (95% CI, 1.18-2.21) for lung cancer, 1.72 (95% CI, 1.18-2.21) for bladder cancer, and 3.09 (95% CI, 1.77-6.43) for melanoma. CONCLUSIONS: We find that patients with CTCL are at increased risk of second malignancies, especially Hodgkin and non-Hodgkin lymphoma, lung cancer, bladder cancer, and melanoma. These findings provide evidence of a population at increased risk of malignancy. Early detection may decrease the morbidity burden of second malignancies, thus providing a strong rationale for prospective screening studies.

Surgical management and lymph-node biopsy of rare malignant cutaneous adnexal carcinomas: a population-based analysis of 7591 patients.

OBJECTIVE: To analyze the prognosis of cutaneous adnexal malignancies, survival relative to surgical management, and utility of lymph-node biopsy. DESIGN: Population-based study of the SEER-18 database from 1975 to 2016. PARTICIPANTS: 7591 patients with sweat gland carcinoma, hidradenocarcinoma, spiradenocarcinoma, sclerosing sweat duct tumor/microcystic adnexal tumor (SSDT/MAC), porocarcinoma, eccrine adenocarcinoma, and sebaceous carcinoma RESULTS: Five-year OS ranged from 68.0 to 82.6%, while 5-year DSS ranged from 94.6 to 99.0%. The majority of patients were treated with narrow (42.4%) or wide local excision (16.9%). DSS at 5 years showed that patients with stage IV had significantly poorer survival (50.3%) than I, II, or III (99.3%, 97.8%, and 89.0% respectively). 5-year OS was significantly higher for narrow excision (excision with < 1 cm margin, 78.5%) than observation (65.0%), excisional biopsy (66.8%), or wide local excision (WLE, 73.2%). Lymph-node biopsy was performed in a minority of cases (8.1%) and patients showed no significant difference in survival based on nodal status. The sensitivity and specificity of lymph-node biopsy for all malignancies were 46% and 80%, respectively. The PPV and NPV for that group were 0.46 and 0.80, respectively. Invasion of deep extradermal structures was a poor predictor of nodal positivity. CONCLUSIONS: These malignancies have excellent DSS. Narrow excisions demonstrate better 5-year DSS and OS compared with WLE. Lymph-node biopsy is a poor predictor of survival in advanced stage disease and utility is limited.

A Nationwide Medical Student Assessment of Oncology Education.

Cancer is the second leading cause of death in the USA, but there is minimal data on how oncology is taught to medical students. The purpose of this study is to characterize oncology education at US medical schools. An electronic survey was sent between December 2014 and February 2015 to a convenience sample of medical students who either attended the American Society for Radiation Oncology annual meeting or serve as delegates to the American Association of Medical Colleges. Information on various aspects of oncology instruction at participants' medical schools was collected. Seventy-six responses from students in 28 states were received. Among the six most common causes of death in the USA, cancer reportedly received the fourth most curricular time. During the first, second, and third years of medical school, participants most commonly reported 6-10, 16-20, and 6-10 h of oncology teaching, respectively. Participants were less confident in their understanding of cancer treatment than workup/diagnosis or basic science/natural history of cancer (p < 0.01). During the preclinical years, pathologists, scientists/Ph.D.'s, and medical oncologists reportedly performed the majority of teaching, whereas during the clinical clerkships, medical and surgical oncologists reportedly performed the majority of teaching. Radiation oncologists were significantly less involved during both periods (p < 0.01). Most schools did not require any oncology-oriented clerkship. During each mandatory rotation, <20 % of patients had a primary diagnosis of cancer. Oncology education is often underemphasized and fragmented with wide variability in content and structure between medical schools, suggesting a need for reform.

Medical student mentorship in radiation oncology at a single academic institution: A 10-year analysis.

PURPOSE: Mentorship has been identified by medical students, residents, and faculty as an important component of specialty selection and research productivity in radiation oncology. This study quantitatively analyzes the impact of a mentorship program in radiation oncology targeted to medical students at our institution. METHODS AND MATERIALS: We performed a retrospective review of 76 current or former medical students who were mentored by faculty radiation oncologists at our institution between 2004 and 2013. Data were collected from the medical school's Office of Student Affairs and from internal departmental records. Mentees were organized by mentorship tracks, which included a clinical track and a research track. For each track, data were compiled and analyzed for student specialty selection, and Fisher exact tests were used to determine the relative significance of exposure to clinical, research, or both tracks on student likelihood of pursuing residency in radiation oncology relative to other specialties. We further tracked the research productivity of mentees in the program, as determined by the number publications that were coauthored by mentees and mentors each year. RESULTS: The absolute number of mentees has grown each year, with a total of 76 mentees, including 58 alumni, at the end of 2013. Mentees in the program have produced a total of 53 manuscripts, given 75 presentations at national conferences, and received numerous national and internal medical school research awards. Of the 58 alumni, 17 (29.3%) applied to and matched into radiation oncology residencies. Alumni of both the research and the clinical track were 5.76 (P < .01) times more likely to enter a radiation oncology residency program than the average single-track alumnus. CONCLUSIONS: Mentorship in medical school is an important factor in the development of future radiation oncologists. These results demonstrate the positive impact mentorship has on specialty selection and research productivity.