Background Over half a million individuals are diagnosed with head and neck cancer each year globally. Radiotherapy is an important curative treatment for this disease, but it requires manual time to delineate radiosensitive organs at risk. This planning process can delay treatment while also introducing interoperator variability, resulting in downstream radiation dose differences. Although auto-segmentation algorithms offer a potentially time-saving solution, the challenges in defining, quantifying, and achieving expert performance remain. Objective Adopting a deep learning approach, we aim to demonstrate a 3D U-Net architecture that achieves expert-level performance in delineating 21 distinct head and neck organs at risk commonly segmented in clinical practice. Methods The model was trained on a data set of 663 deidentified computed tomography scans acquired in routine clinical practice and with both segmentations taken from clinical practice and segmentations created by experienced radiographers as part of this research, all in accordance with consensus organ at risk definitions. Results We demonstrated the model’s clinical applicability by assessing its performance on a test set of 21 computed tomography scans from clinical practice, each with 21 organs at risk segmented by 2 independent experts. We also introduced surface Dice similarity coefficient, a new metric for the comparison of organ delineation, to quantify the deviation between organ at risk surface contours rather than volumes, better reflecting the clinical task of correcting errors in automated organ segmentations. The model’s generalizability was then demonstrated on 2 distinct open-source data sets, reflecting different centers and countries to model training. Conclusions Deep learning is an effective and clinically applicable technique for the segmentation of the head and neck anatomy for radiotherapy. With appropriate validation studies and regulatory approvals, this system could improve the efficiency, consistency, and safety of radiotherapy pathways.

Purpose: This review provides an update on the current clinical, epidemiological and pathophysiological evidence alongside the diagnostic, prevention and treatment approach to chemotherapy-induced peripheral neuropathy (CIPN). Findings: The incidence of cancer and longterm survival after treatment is increasing. CIPN affects sensory, motor and autonomic nerves and is one of the most common adverse events caused by chemotherapeutic agents, which in severe cases leads to dose reduction or

BACKGROUND Over half a million individuals are diagnosed with head and neck cancer each year globally. Radiotherapy is an important curative treatment for this disease, but it requires manual time to delineate radiosensitive organs at risk. This planning process can delay treatment while also introducing interoperator variability, resulting in downstream radiation dose differences. Although auto-segmentation algorithms offer a potentially time-saving solution, the challenges in defining, quantifying, and achieving expert performance remain. OBJECTIVE Adopting a deep learning approach, we aim to demonstrate a 3D U-Net architecture that achieves expert-level performance in delineating 21 distinct head and neck organs at risk commonly segmented in clinical practice. METHODS The model was trained on a data set of 663 deidentified computed tomography scans acquired in routine clinical practice and with both segmentations taken from clinical practice and segmentations created by experienced radiographers as part of this research, all in accordance with consensus organ at risk definitions. RESULTS We demonstrated the model’s clinical applicability by assessing its performance on a test set of 21 computed tomography scans from clinical practice, each with 21 organs at risk segmented by 2 independent experts. We also introduced surface Dice similarity coefficient, a new metric for the comparison of organ delineation, to quantify the deviation between organ at risk surface contours rather than volumes, better reflecting the clinical task of correcting errors in automated organ segmentations. The model’s generalizability was then demonstrated on 2 distinct open-source data sets, reflecting different centers and countries to model training. CONCLUSIONS Deep learning is an effective and clinically applicable technique for the segmentation of the head and neck anatomy for radiotherapy. With appropriate validation studies and regulatory approvals, this system could improve the efficiency, consistency, and safety of radiotherapy pathways.

The fields of radiotherapy and clinical oncology have been rapidly changed by the advances of technology. Improvement in computer processing power and imaging quality heralded precision radiotherapy allowing radiotherapy to be delivered efficiently, safely and effectively for patient benefit. Artificial intelligence (AI) is an emerging field of computer science which uses computer models and algorithms to replicate human-like intelligence and perform specific tasks which offers a huge potential to healthcare. We reviewed and presented the history, evolution and advancement in the fields of radiotherapy, clinical oncology and machine learning. Radiotherapy target delineation is a complex task of outlining tumour and organ at risks volumes to allow accurate delivery of radiotherapy. We discussed the radiotherapy planning, treatment delivery and reviewed how technology can help with this challenging process. We explored the evidence and clinical application of machine learning to radiotherapy. We concluded on the challenges, possible future directions and potential collaborations to achieve better outcome for cancer patients.

6508 Background: Most newly diagnosed oro- & hypopharngeal cancers (OPC, HPC) are treated with (chemo)RT with curative intent but at the consequence of adverse effects on quality of life. CRUK/14/014 investigated if using Do-IMRT to reduce RT dose to the dysphagia/aspiration related structures (DARS) improved swallowing function compared to S-IMRT. Methods: Patients with T1-4, N0-3, M0 OPC/HPC were randomised 1:1 to S-IMRT (65 Gray (Gy)/30 fractions (f) to primary & nodal tumour; 54Gy/30f to remaining pharyngeal subsite & nodal areas at risk of microscopic disease) or Do-IMRT. The volume of the superior & middle pharyngeal constrictor muscle (PCM) (OPC) or inferior PCM (HPC) lying outside the high-dose target volume was set a mandatory mean dose constraint in Do-IMRT. Treatment allocation was by minimisation balanced by centre, use of induction/concomitant chemotherapy, tumour site & AJCC stage. Primary endpoint was mean MD Anderson Dysphagia Inventory (MDADI) composite score 12 months after RT with 102 patients needed to detect a 10 point improvement (assuming S-IMRT score of 72, standard deviation (SD) 13.8; 90% power, 2-sided 5% alpha). Patients were blind to treatment allocation. Secondary endpoints included local control. Results: 112 patients (56 S-IMRT, 56 Do-IMRT) were randomised from 22 UK centres from 06/2016 to 04/2018. Mean age was 57 years; 80% were male; 97% had OPC; 90% had AJCC stage 3&4 disease; 86% had concomitant chemotherapy only, 4% induction & concomitant and 10% no chemotherapy. 111/112 had RT doses as prescribed (1 patient died before RT). Median of the mean inferior PCM dose was S-IMRT 49.8Gy (IQR 47.1-52.4) vs. Do-IMRT 28.4Gy (21.3–37.4), p < 0.0001; superior & middle PCM dose was S-IMRT 57.2Gy (56.3–58.3) vs. Do-IMRT 49.7Gy (49.4–49.9), p < 0.0001. Do-IMRT had significantly higher MDADI scores: S-IMRT 70.3 (SD 17.3) vs. Do-IMRT 77.7 (16.1), p = 0.016. 3 local recurrences (1 S-IMRT, 2 Do-IMRT) have been reported. Conclusions: Do-IMRT reduced RT dose to the DARS and improved patient reported swallowing function compared with S-IMRT. This is the first randomised study to demonstrate functional benefit of swallow-sparing IMRT in OPC. Clinical trial information: 25458988 .