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.

Development and internal validation of multivariable prediction models for biochemical failure after MRI-guided focal salvage high-dose-rate brachytherapy for radiorecurrent prostate cancer.

BACKGROUND AND PURPOSE: Magnetic resonance-guided focal salvage high-dose-rate brachytherapy (FS-HDR-BT) for radiorecurrent prostate cancer (PCa) shows low toxicity rates. However, biochemical failure (BF) after treatment occurs frequently. We developed two prediction models for BF (Phoenix definition) with the aim of enhancing patient counselling before FS-HDR-BT and during follow-up. MATERIALS AND METHODS: A prospective cohort of 150 radiorecurrent PCa patients treated with FS-HDR-BT between 2013 and 2020 was used for model development and internal validation. Multivariable Cox Proportional Hazards regression was applied. For model 1, only pre-salvage variables were included as candidate predictors. For model 2, additional (post-)salvage characteristics were tested. After calibration, nomograms and webtools were constructed. Finally, three risk groups were identified. RESULTS: Sixty-one patients (41%) experienced BF. At baseline (model 1), age, gross tumour volume, pre-salvage PSA, and pre-salvage PSA doubling time (PSADT) were predictive of BF. During follow-up (model 2), age, pre-salvage PSA and PSADT, seminal vesicle involvement, post-salvage time to PSA nadir, and percentage PSA reduction were predictive of BF. The adjusted C-statistics were 0.73 (95% CI: 0.66-0.81) and 0.84 (95% CI: 0.78-0.90), respectively, with acceptable calibration. Estimated 2-year biochemical disease-free survival for the low-, intermediate-, and high-risk groups were 84%, 70%, and 31% (model 1), and 100%, 71%, and 5% (model 2). CONCLUSION: Two models are provided for prediction of BF in patients with radiorecurrent PCa treated with FS-HDR-BT. Based on pre- and post-salvage characteristics, we are able to identify patients with a high risk of BF. These findings can aid patient counselling for FS-HDR-BT.

Dosimetric impact of contouring and image registration variability on dynamic 125I prostate brachytherapy.

PURPOSE: The quality of permanent prostate brachytherapy can be increased by addition of imaging modalities in the intraoperative procedure. This addition involves image registration, which inherently has inter- and intraobserver variabilities. We sought to quantify the inter- and intraobserver variabilities in geometry and dosimetry for contouring and image registration and analyze the results for our dynamic 125I brachytherapy procedure. METHODS AND MATERIALS: Five observers contoured 11 transrectal ultrasound (TRUS) data sets three times and 11 CT data sets one time. The observers registered 11 TRUS and MRI data sets to cone beam CT (CBCT) using fiducial gold markers. Geometrical and dosimetrical inter- and intraobserver variabilities were assessed. For the contouring study, structures were subdivided into three parts along the craniocaudal axis. RESULTS: We analyzed 165 observations. Interobserver geometrical variability for prostate was 1.1 mm, resulting in a dosimetric variability of 1.6% for V100 and 9.3% for D90. The geometric intraobserver variability was 0.6 mm with a V100 of 0.7% and D90 of 1.1%. TRUS-CBCT registration showed an interobserver variability in V100 of 2.0% and D90 of 3.1%. Intraobserver variabilities were 0.9% and 1.6%, respectively. For MRI-CBCT registration, V100 and D90 were 1.3% and 2.1%. Intraobserver variabilities were 0.7% and 1.1% for the same. CONCLUSIONS: Prostate dosimetry is affected by interobserver contouring and registration variability. The observed variability is smaller than underdosages that are adapted during our dynamic brachytherapy procedure.

Cone-beam CT-based adaptive planning improves permanent prostate brachytherapy dosimetry: An analysis of 1266 patients.

PURPOSE: To evaluate adaptive planning for permanent prostate brachytherapy and to identify the prostate regions that needed adaptation. METHODS AND MATERIALS: After the implantation of stranded seeds, using real-time intraoperative planning, a transrectal ultrasound (TRUS)-scan was obtained and contoured. The positions of seeds were determined on a C-arm cone-beam computed tomography (CBCT)-scan. The CBCT-scan was registered to the TRUS-scan using fiducial gold markers. If dose coverage on the combined image-dataset was inadequate, an intraoperative adaptation was performed by placing remedial seeds. CBCT-based intraoperative dosimetry was analyzed for the prostate (D90, V100, and V150) and the urethra (D30). The effects of the adaptive dosimetry procedure for Day 30 were separately assessed. RESULTS: We analyzed 1266 patients. In 17.4% of the procedures, an adaptation was performed. Without the dose contribution of the adaptation Day 30 V100 would be < 95% for half of this group. On Day 0, the increase due to the adaptation was 11.8 ± 7.2% (1SD) for D90 and 9.0 ± 6.4% for V100. On Day 30, we observed an increase in D90 of 12.3 ± 6.0% and in V100 of 4.2 ± 4.3%. For the total group, a D90 of 119.6 ± 9.1% and V100 of 97.7 ± 2.5% was achieved. Most remedial seeds were placed anteriorly near the base of the prostate. CONCLUSION: CBCT-based adaptive planning enables identification of implants needing adaptation and improves prostate dose coverage. Adaptations were predominantly performed near the anterior base of the prostate.

Rectal dose constraints for salvage iodine-125 prostate brachytherapy.

PURPOSE: Organ-confined prostate cancer recurrences after primary radiotherapy can be treated with salvage iodine-125 brachytherapy. Options include total salvage (TS) or focal salvage (FS). TS often leads to severe late gastrointestinal (GI) toxicity. Differences in rectal dosimetry between TS and FS are presented and dose constraints proposed to reduce late severe GI toxicity (>90 days). METHODS AND MATERIALS: Intraoperative dosimetry and 30-day CT-dosimetry of 20 FS and 28 TS patients were evaluated. GI toxicity was evaluated using the common terminology criteria for adverse events-4. With receiver operating characteristic analysis, dosimetry cutoff values to prevent severe late GI toxicity were assessed. RESULTS: FS reduces rectal dose significantly. Median D(0.1cc), D(1cc), D(2cc), and V100 reductions were 38 Gy (p = 0.002), 46 Gy (p < 0.0001), 46 Gy (p < 0.0001), and 0.41 cc (p = 0.0001), respectively, compared with TS. FS patients had no late severe GI toxicity. TS patients with severe GI toxicity (41%, n = 11) showed significantly higher rectal doses than TS patients without GI toxicity (59%, n = 16). Median D(0.1cc), D(1cc), D(2cc), and V100 differences were 29 Gy (p < 0.001), 17 Gy (p = 0.001), 28 Gy (p < 0.001), and 0.45 cc (p = 0.001). With receiver operating characteristic analysis, restrictions for the D(0.1cc), D(1cc), D(2cc), and V100 are <160 Gy (area under the curve [AUC], 0.88; 95% confidence interval [CI] 0.76-1.00), <119 Gy (AUC, 0.87; 95% CI, 0.74-1.00), <102 Gy (AUC, 0.89; 95% CI, 0.77-1.00), and <0.38 cc (AUC, 0.88; 95% CI, 0.75-1.00), respectively. Thirty-day CT dosimetry showed minor overestimation of intraoperative D(2cc) (median, 10 Gy [p = 0.02]). CONCLUSIONS: FS reduces rectal dose compared with TS. D(0.1cc), D(1cc), D(2cc), and V100 restrictions were 160 Gy, 120 Gy, 100 Gy, and 0.35 cc. Taking correlation into account, the D2cc <100 Gy might be sufficient for clinical practice. Larger series and multivariable models are necessary to further assess the found restrictions.

The significance of anemia in squamous cell head and neck cancer treated with surgery and postoperative radiotherapy.

The objective of this study was to investigate the prognostic significance of the hemoglobin (Hb) levels at different timepoints in locally advanced squamous cell carcinoma of the head and neck (SCCHN). Included were 111 patients. The hemoglobin levels were assessed before surgery (PreS-Hb), between surgery and radiotherapy (HbAAC), before postoperative radiotherapy (PreRT-Hb) and at the end of radiotherapy (EndRT-Hb). HbAAC takes into account the duration of anemia during the interval between surgery and radiotherapy. Higher HbAAC corresponds with lower Hb levels. Five year locoregional control (LRC) among patients with HbAAC > or = median was 72% and significantly worse as compared to the 88% in case of HbAAC < median (p = 0.0097). Multivariate analysis for LRC showed that the HbAAC was a prognostic factor. Overall survival (OS) after 3 years was 77% in case of HbAAC < median and 34% in case of HbAAC > or = median (p < 0.0002). Multivariate analysis for OS showed that the PreS-Hb and HbAAC were prognostic factors. Hb level between surgery and radiotherapy is an important prognostic factor for both LRC and OS among patients with SCCHN treated with surgery and postoperative radiotherapy.