Deep learning for automated contouring of neurovascular structures on magnetic resonance imaging for prostate cancer patients.

Background and purpose: Manual contouring of neurovascular structures on prostate magnetic resonance imaging (MRI) is labor-intensive and prone to considerable interrater disagreement. Our aim is to contour neurovascular structures automatically on prostate MRI by deep learning (DL) to improve workflow and interrater agreement. Materials and methods: Segmentation of neurovascular structures was performed on pre-treatment 3.0 T MRI data of 131 prostate cancer patients (training [n = 105] and testing [n = 26]). The neurovascular structures include the penile bulb (PB), corpora cavernosa (CCs), internal pudendal arteries (IPAs), and neurovascular bundles (NVBs). Two DL networks, nnU-Net and DeepMedic, were trained for auto-contouring on prostate MRI and evaluated using volumetric Dice similarity coefficient (DSC), mean surface distances (MSD), Hausdorff distances, and surface DSC. Three radiation oncologists evaluated the DL-generated contours and performed corrections when necessary. Interrater agreement was assessed and the time required for manual correction was recorded. Results: nnU-Net achieved a median DSC of 0.92 (IQR: 0.90-0.93) for the PB, 0.90 (IQR: 0.86-0.92) for the CCs, 0.79 (IQR: 0.77-0.83) for the IPAs, and 0.77 (IQR: 0.72-0.81) for the NVBs, which outperformed DeepMedic for each structure (p < 0.03). nnU-Net showed a median MSD of 0.24 mm for the IPAs and 0.71 mm for the NVBs. The median interrater DSC ranged from 0.93 to 1.00, with the majority of cases (68.9%) requiring manual correction times under two minutes. Conclusions: DL enables reliable auto-contouring of neurovascular structures on pre-treatment MRI data, easing the clinical workflow in neurovascular-sparing MR-guided radiotherapy.

Recurrence characteristics after focal salvage HDR brachytherapy in prostate cancer.

BACKGROUND AND PURPOSE: Radiorecurrent prostate cancer is often confined to the prostate, predominantly near the index lesion. The purpose of this study was to look at recurrence characteristics in patients treated with focal salvage high dose-rate (HDR) brachytherapy. MATERIALS AND METHODS: Patients treated with MRI-guided HDR brachytherapy, with a single fraction of 19 Gy from July 2013 to October 2021 as focal salvage treatment, were prospectively included in the current study. Imaging data were collected regarding the occurrence of local, regional and distant recurrences, including location of local recurrences (LR) in relation to the HDR radiotherapy field. RESULTS: One hundred seventy-five patients were included after focal salvage HDR brachytherapy (median follow-up 36 months (IQR 23-50)). Three-years biochemical recurrence-free survival, LR-free survival, in-field LR-free survival, out-of-field LR-free survival, any-recurrence-free survival and ADT-free survival were 43% (95%CI 34%-52%), 51% (41%-61%), 70% (61%-80%), 92% (88%-97%), 42% (32%-52%) and 86% (80%-92%), respectively. Larger GTV-size and shorter PSA doubling time were associated with in-field LR in multivariable analysis. CONCLUSION: After focal salvage HDR brachytherapy with a dose of 1x19 Gy for local prostate cancer recurrence, subsequent recurrences are mostly local and in-field.

[Experiences with MRI-guided brachytherapy for radiorecurrent prostate cancer]. / Ervaringen met MRI-geleide brachytherapie bij prostaatkankerrecidief.

MR-guided brachytherapy offers a focal salvage treatment for the local recurrence in case of isolated locally recurrent prostate cancer in the prostate and/or seminal vesicles after primary radiotherapy. By focusing on only the local recurrence instead of the whole prostate, chances of additional toxicity of the bladder, urethra and rectum can be minimized. In almost all patients, the treatment leads to a good initial treatment response that persists in about half of patients, while others will develop progressive disease later on. For selecting suitable patients, factors such as preexistent urinary- and bowel complaints, localization and size of the recurrence, PSA doubling time and time between primary radiotherapy and development of the recurrence are relevant. MR-guided brachytherapy can provide a suitable salvage strategy, with the aims of deferring androgen deprivation therapy and a chance of cure.

Online adaptive MR-guided radiotherapy: Conformity of contour adaptation for prostate cancer, rectal cancer and lymph node oligometastases among radiation therapists and radiation oncologists.

Background and purpose: Online adaptive MR-guided treatment planning workflows facilitate daily contour adaptation to the actual anatomy. Allocating contour adaptation to radiation therapists (RTTs) instead of radiation oncologists (ROs) might allow for increasing workflow efficiency. This study investigates conformity of adapted target contours provided by dedicated RTTs and ROs. Materials and methods: In a simulated online procedure, 6 RTTs and 6 ROs recontoured targets and organs at risk (OAR) in prostate cancer (n = 2), rectal cancer (n = 2) and lymph node-oligometastases (n = 2) cases. RTTs gained contouring competence beforehand by following a specific in-house training program. For all target contours and the reference delineations volumetric differences were determined and Dice similarity coefficient (DSC), conformity index (CI) and generalized CI were calculated. Delineation time and -confidence were registered for targets and OAR. Impact of contour adaptation on treatment plan quality was investigated. Results: Delineation conformity was generally high with DSC, CI and generalized CI values in the range of 0.81-0.94, 0.87-0.95 and 0.63-0.85 for prostate cancer, rectal cancer and LN-oligometastasis, respectively. Target volumes were comparable for both, RTTs and ROs. Time needed and confidence in contour adaptation was comparable as well. Treatment plans derived with adapted contours did not violate dose volume constrains as used in clinical routine. Conclusion: After tumor site specific training, daily contour adaptations as needed in adaptive online radiotherapy workflows can be accurately performed by RTTs. Conformity of the derived contours is high and comparable to contours as provided by ROs.

Non-Small-Cell Lung Cancer Patients with a High Predicted Risk of Irradical Resection: Can Chemoradiotherapy Offer Similar Survival?

PURPOSE: Irradical resection of non-small-cell lung cancer (NSCLC) is a detrimental prognostic factor. Recently, Rasing et al. presented an internationally validated risk score for pre-treatment prediction of irradical resection. We hypothesized that chemoradiation therapy (CRT) could serve as an alternative approach in patients with a high risk score and compared overall survival (OS) outcomes between surgery and CRT. METHODS: Patients from a population-based cohort with stage IIB-III NSCLC between 2015 and 2018 in The Netherlands were selected. Patients with a 'Rasing score' > 4 who underwent surgery were matched with patients who underwent CRT using 1:1 nearest-neighbor propensity score matching. The primary endpoint of OS was compared using a Kaplan-Meier analysis. RESULTS: In total, 2582 CRT and 638 surgery patients were eligible. After matching, 523 well-balanced pairs remained. Median OS in the CRT group was 27.5 months, compared with 45.6 months in the surgery group (HR 1.44, 95% CI 1.23-1.70, p < 0.001). The 114 surgical patients who underwent an R1-2 resection (21.8%) had a worse median OS than the CRT group (20.2 versus 27.5 months, HR 0.77, 95% CI 0.61-0.99, p = 0.039). CONCLUSION: In NSCLC patients at high predicted risk of irradical resection, CRT appears to yield inferior survival compared with surgery. Therefore, choosing CRT instead of surgery cannot solely be based on the Rasing score. Since patients receiving an R1-2 resection do have detrimental outcomes compared with primary CRT, the treatment decision should be based on additional information, such as imaging features, comorbidities, patient preference, and the surgeon's confidence in achieving an R0 resection.

Adjuvant Treatment Following Irradical Resection of Stage I-III Non-small Cell Lung Cancer: A Population-based Study.

Irradical (R1-2) resection for non-small cell lung cancer (NSCLC) is associated with a dismal prognosis. Adjuvant treatment attempts to improve survival outcomes, but evidence on the optimal strategy is limited. The purpose of this study was to compare overall survival (OS) between different adjuvant treatment strategies in these patients. Out of 8,528 patients with newly diagnosed NSCLC from 2015-2018, those with an R1-2 resection were identified from the Netherlands Cancer Registry. First, OS was compared between adjuvant treatment groups 'no therapy', 'radiotherapy (RT) only', 'chemotherapy only', and 'chemo- and radiotherapy (CRT)' using multinomial propensity score-weighted Cox regression analysis. Second, three 1:1 propensity score-matched sets were created for chemotherapy vs no chemotherapy, RT only vs no therapy, and CRT vs chemotherapy only. Kaplan-Meier and Cox regression analyses for OS were performed in each set. With a median follow-up of 23 months, 427 patients were selected. In the weighted regression analysis, compared to no adjuvant therapy, chemotherapy and CRT were associated with improved OS (HR 0.41, 95%CI: 0.22-0.76; and HR 0.55, 95%CI: 0.37-0.81, respectively), whereas RT was not (HR 1.04, 95%CI: 0.73-1.50). In the matched sets, OS was improved after chemotherapy (+/- RT) compared to no chemotherapy (HR 0.47, 95%CI: 0.32-0.69). No OS difference was observed between matched groups of RT only vs no adjuvant therapy (HR 1.13, 95%CI: 0.74-1.72), nor for CRT vs chemotherapy only (HR 1.37, 95%CI: 0.70-2.71). Adjuvant chemotherapy, but not radiotherapy, improves survival after an R1-2 resection in stage I-III NSCLC.

Predicting Incomplete Resection in Non-Small Cell Lung Cancer Preoperatively: A Validated Nomogram.

BACKGROUND: Patients who are surgically treated for stage I to III non-small cell lung cancer (NSCLC) have dismal prognosis after incomplete (R1-R2) resection. Our study aimed to develop a prediction model to estimate the chance of incomplete resection based on preoperative patient-, tumor-, and treatment-related factors. METHODS: From a Dutch national cancer database, NSCLC patients who had surgical treatment without neoadjuvant therapy were selected. Thirteen possible predictors were analyzed. Multivariable logistic regression was used to create a prediction model. External validation was applied in the American National Cancer Database, whereupon the model was adjusted. Discriminatory ability and calibration of the model was determined after internal and external validation. The prediction model was presented as nomogram. RESULTS: Of 7156 patients, 511 had an incomplete resection (7.1%). Independent predictors were histology, cT stage, cN stage, extent of surgery, and open vs thoracoscopic approach. After internal validation, the corrected C statistic of the resulting nomogram was 0.72. Application of the nomogram to an external data set of 85,235 patients with incomplete resection in 2485 patients (2.9%) resulted in a C statistic of 0.71. Calibration revealed good overall fit of the nomogram in both cohorts. CONCLUSIONS: An internationally validated nomogram is presented providing the ability to predict the individual chance of incomplete resection in patients with stage I to III NSCLC planned for resection. In case of a high predicted risk of incomplete resection, alternative treatment strategies could be considered, whereas a low risk further supports the use of surgical procedures.