The recent rise of deep learning (DL) and its promising capabilities in capturing non-explicit detail from large datasets have attracted substantial research attention in the field of medical image processing. DL provides grounds for technological development of computer-aided diagnosis and segmentation in radiology and radiation oncology. Amongst the anatomical locations where recent auto-segmentation algorithms have been employed, the pelvis remains one of the most challenging due to large intra- and inter-patient soft-tissue variabilities. This review provides a comprehensive, non-systematic and clinically-oriented overview of 74 DL-based segmentation studies, published between January 2016 and December 2020, for bladder, prostate, cervical and rectal cancers on computed tomography (CT) and magnetic resonance imaging (MRI), highlighting the key findings, challenges and limitations.

The implementation of plan of the day selection for patients receiving radiotherapy (RT) for bladder cancer requires efficient and confident decision-making. This article describes the development of a training programme and maintenance of competency. Cone beam CT (CBCT) images acquired on patients receiving RT for bladder cancer were assessed to establish baseline competency and training needs. A training programme was implemented, and observers were asked to select planning target volumes (PTVs) on two groups of 20 patients' images. After clinical implementation, the PTVs chosen were reviewed offline, and an audit performed after 3 years. A mean of 73% (range, 53-93%) concordance rate was achieved prior to training. Subsequent to training, the mean score decreased to 66% (Round 1), then increased to 76% (Round 2). Six radiographers and two clinicians successfully completed the training programme. An independent observer reviewed the images offline after clinical implementation, and a 91% (126/139) concordance rate was achieved. During the audit, 125 CBCT images from 13 patients were reviewed by a single observer and concordance was 92%. Radiographer-led selection of plan of the day was implemented successfully with the use of a training programme and continual assessment. Quality has been maintained over a period of 3 years. The training programme was successful in achieving and maintaining competency for a plan of the day technique.

IGRT in cervical cancer treatment delivery is complex due to significant target and organs at risk (OAR) motion. Implementing image assessment of soft-tissue target and OAR position to improve accuracy is recommended. We report the development and refinement of a training and competency programme (TCP), leading to on-line Radiation Therapist (RTT) led soft-tissue assessment, evaluated by a prospective audit. The TCP comprised didactic lectures and practical sessions, supported by a comprehensive workbook. The content was decided by a team comprised of Clinical Oncologists, RTTs, and Physicists. On completion of training, RTT soft-tissue review proficiency (after bony anatomy registration) was assessed against a clinician gold-standard from a database of 20 cervical cancer CBCT images. Reviews were graded pass or fail based on PTV coverage assessment and decision taken in concordance with the gold-standard. Parity was set at ≥80% agreement.The initial TCP (stage one) focussed on offline verification and decision making. Sixteen RTTs completed this stage, four achieved ≥80%. This was not sufficient to support clinical implementation.The TCP was redesigned, more stringent review guidelines and greater anatomy teaching was added. TCP stage two focussed on online verification and decision making supported by a decision flowchart. Twenty-one RTTs completed this TCP, all achieved ≥80%. This supported clinical implementation of RTT-led soft-tissue review under prospective audit conditions.The prospective audit was conducted between March 2017 and August 2017. Daily online review was performed by two trained RTTs. Online review and decision making proficiency was evaluated by a clinician. Thirteen patients were included in the audit. Daily online RTT-led IGRT was achieved for all 343 fractions. Two-hundred CBCT images were reviewed offline by the clinician; the mean number of reviews per patient was 15. 192/200 (96%) RTT image reviews were in agreement with clinician review, presenting excellent concordance. Multidisciplinary involvement in training development, redesign of the TCP and inclusion of summative competency assessment were important factors to support RTT skill development. Consequently, RTT-led cervical cancer soft-tissue IGRT was clinically implemented in the hospital.

In ENGOT-cx11/GOG-3047/KEYNOTE-A18 (NCT04221945), pembrolizumab (vs placebo) + concurrent chemoradiotherapy (CCRT) followed by pembrolizumab (vs placebo) significantly improved progression-free survival and overall survival in participants with newly diagnosed, high-risk locally advanced cervical cancer (LACC). We report patient-reported outcomes (PROs) from the study. Participants (≥18 years) with high-risk LACC (FIGO 2014 stage IB2-IIB with node-positive disease or stage III-IVA) were randomized 1:1 to 5 cycles of pembrolizumab 200 mg or placebo Q3W plus CCRT, followed by 15 cycles of pembrolizumab 400 mg or placebo Q6W. CCRT was 5 cycles (optional 6th dose) of cisplatin 40 mg/m² Q1W plus external beam radiotherapy followed by brachytherapy. Secondary PRO endpoints included EORTC QLQ-C30 GHS/QoL and physical functioning subscales and EORTC QLQ-CX24 symptom experience scale; EQ-5D-5L visual analogue scale (VAS) was an exploratory endpoint. No alpha was assigned to the PRO analyses. 1008 (95.1 %) of 1060 randomized participants were included in the PRO full analysis set population. No meaningful between-group differences were observed for any of the prespecified PRO instruments. Between-group differences (95 % CIs) in least-squares mean score changes from baseline to week 36 (primary time point) were 0.57 (-2.34 to 3.49) for QLQ-C30 GHS/QoL, 0.64 (-1.54 to 2.82) for QLQ-C30 physical functioning, 0.54 (-1.02 to 2.09) for QLQ-CX24 symptom experience, and -0.55 (-2.97 to 1.86) for EQ-5D-5L VAS. Empirical mean score changes over time and the proportions of participants whose scores improved, remained stable, or deteriorated over time were similar between treatment arms. Pembrolizumab+CCRT did not negatively impact participants' health-related quality of life.

Objectives: Quantify target volume delineation uncertainty for CT/MRI simulation and MRI-guided adaptive radiotherapy in rectal cancer. Define optimal imaging sequences for target delineation. Methods: Six experienced radiation oncologists delineated clinical target volumes (CTVs) on CT and 2D and 3D-MRI in three patients with rectal cancer, using consensus contouring guidelines. Tumour GTV (GTVp) was also contoured on MRI acquired week 0 and 3 of radiotherapy. A STAPLE contour was created and volume and interobserver variability metrics were analysed. Results: There were statistically significant differences in volume between observers for CT and 2D-MRI-defined CTVs (p < 0.05). There was no significant difference between observers on 3D-MRI. Significant differences in volume were seen between observers for both 2D and 3D-MRI-defined GTVp at weeks 0 and 3 (p < 0.05). Good interobserver agreement (IOA) was seen for CTVs delineated on all imaging modalities with best IOA on 3D-MRI; median Conformity index (CI) 0.74 for CT, 0.75 for 2D-MRI and 0.77 for 3D-MRI. IOA of MRI-defined GTVp week 0 was better compared to CT; CI 0.58 for CT, 0.62 for 2D-MRI and 0.7 for 3D-MRI. MRI-defined GTVp IOA week three was worse compared to week 0. Conclusion: Delineation on MRI results in smaller volumes and better IOA week 0 compared to CT. 3D-MRI provides the best IOA in CTV and GTVp. MRI-defined GTVp on images acquired week 3 showed worse IOA compared to week 0. This highlights the need for consensus guidelines in GTVp delineation on MRI during treatment course in the context of dose escalation MRI-guided rectal boost studies. Advances in knowledge: Optimal MRI sequences for CT/MRI simulation and MRI-guided adaptive radiotherapy in rectal cancer have been defined.

Background: The majority of locally advanced cervical cancers (LaCC) are causally related to HPV. We sought to investigate the utility of an ultra-sensitive HPV-DNA next generation sequencing (NGS) assay—panHPV-detect—in LaCC treated with chemoradiotherapy, as a marker of treatment response and persistent disease. Method: Serial blood samples were collected from 22 patients with LaCC before, during and after chemoradiation. The presence of circulating HPV-DNA was correlated with clinical and radiological outcomes. Results: The panHPV-detect test demonstrated a sensitivity and specificity of 88% (95% CI-70–99%) and 100% (95% CI-30–100%), respectively, and correctly identified the HPV-subtype (16, 18, 45, 58). After a median follow up of 16 months, and three relapses all had detectable cHPV-DNA at 3 months post-CRT despite complete response on imaging. Another four patients with radiological partial or equivocal response and undetectable cHPV-DNA at the 3-month time point did not go on to develop relapse. All patients with radiological CR and undetectable cHPV-DNA at 3-months remained disease free. Conclusions: These results demonstrate that the panHPV-detect test shows high sensitivity and specificity for detecting cHPV-DNA in plasma. The test has potential applications in assessment of the response to CRT and in monitoring for relapse, and these initial findings warrant validation in a larger cohort.