Search

Deep-learning-based segmentation tools have yielded higher reported segmentation accuracies for many medical imaging applications. However, inter-site variability in image properties can challenge the translation of these tools to data from ‘unseen’ sites not included in the training data. This study quantifies the impact of inter-site variability on the accuracy of deep-learning-based segmentations of the prostate from magnetic resonance (MR) images, and evaluates two strategies for mitigating the reduced accuracy for data from unseen sites: training on multi-site data and training with limited additional data from the unseen site. Using 376 T2-weighted prostate MR images from six sites, we compare the segmentation accuracy (Dice score and boundary distance) of three deep-learning-based networks trained on data from a single site and on various configurations of data from multiple sites. We found that the segmentation accuracy of a single-site network was substantially worse on data from unseen sites than on data from the training site. Training on multi-site data yielded marginally improved accuracy and robustness. However, including as few as 8 subjects from the unseen site, e.g. during commissioning of a new clinical system, yielded substantial improvement (regaining 75% of the difference in Dice score).

Academic and professional resistance to focal prostate therapy has been dominated by the fact that it is a multifocal disease. As such, until focal therapy became a legitimate research question, no attempt was made to distinguish between those men with unifocal cancer and those with multifocal disease since this information had little clinical utility in decision-making for therapy. There is growing evidence that the risk of disease progression to local invasion, metastases, and death is likely to be monoclonal in origin with the largest dominant, or so-called index lesion, responsible for the vast majority of these. This chapter reviews the considerable body of evidence that has grown to support this hypothesis.

Template prostate mapping biopsies are a highly accurate investigation of prostate cancer that can be applied to most men who are at risk of prostate cancer. They provide spatial information in the location and extent of diseased tissue which cannot be obtained via TRUS biopsy while confirming absence of the disease in the remaining prostate tissue. They are characteristically better in detection of anteriorly located tumors. It is also an ideal reference standard for validation studies.

Prostate cancer is the most common cancer in males and the second-leading cancer-related cause of death in the Western world. Prostate-specific antigen (PSA) screening, and changes in the diagnostic pathway of prostate cancer, has resulted in proportionally more men being diagnosed with early stage prostate cancer still confined to the prostate. Patients are faced with an overwhelming choice of treatment options to treat their disease. Standard options include radical prostatectomy (RP) and external-beam radiotherapy (ERBT) which, although provide good long-term cancer control, are associated with significant side effects affecting the quality of life of these patients.

New surgical techniques are continuously being developed in the field of prostate cancer, more so with the ever-increasing interest in minimally invasive techniques to treat solid organ cancers. This has been triggered by the current state of play with treating a disease that has a long natural history in which the benefits and risks of radical therapy are not quite right. In other words, whole-gland radical therapy or radiotherapy can cause significant complications that are a direct result of damage to surrounding structures, including erectile dysfunction (30–70 %), urinary incontinence (5–20 %) and bowel toxicity (5–10 %). Focal therapy aims to reduce the complication profile by focusing the therapy to the cancer lesion and preserving surrounding tissues, thus improving functional outcome.