This study will evaluate radiation medicine professionals’ perceptions of clinical and professional risks and benefits, and the evolving roles and responsibilities with artificial intelligence (AI).

Radiation oncologists (ROs), medical physicists (MPs), treatment planners (TP-RTTs) and treatment delivery radiation therapists (TD-RTTs) at a cancer centre in preliminary stages of implementing an AI-enabled treatment planning system were invited to participate in uniprofessional focus groups. Semi-structured scripts addressed the perceptions of AI, including thoughts regarding changing roles and competencies. Sessions were audiorecorded, transcribed and coded thematically through consensus-building.

A total of 24 participants (four ROs, five MPs, seven TP-RTTs and eight TD-RTTs) were engaged in four focus groups of 58 minutes average duration (range 54–61 minutes). Emergent themes addressed AI’s impact on quality of care, changing professional tasks and changing competency requirements. Time-consuming repetitive tasks such as delineating targets, generating treatment plans and quality assurance were thought conducive to offloading to AI. Outcomes data and adaptive planning would be incorporated into clinical decision-making. Changing workload would necessitate changing skills, prioritising plan evaluation over generation and increasing interprofessional communication. All groups discussed AI reducing the need for TP-RTTs, though displacement was thought more likely than replacement.

It is important to consider how professionals perceive AI to be proactive in informing change, as gains in quality and efficiency will require new workflows, skills and education.

The authors proposed a novel perimeter-based index (PBI) that was capable of evaluating the accuracy in the appraisal of auto-segmentation software. A quantitative value, that is time saved in editing the auto-segmented contours, was used to compare the effectiveness of two other commonly used indices in this study.

The relationship between the proposed index and the amount of the contouring time that could be saved was studied. The performances of two other commonly used similarity indices, namely Dice similarity coefficient (DSC) and the modified normalised average Hausdorff distance (MNAHD), were also evaluated. Ten nasopharyngeal cases and ten prostate cases that were previously treated with intensity-modulated radiation therapy technique were recruited as the validation cases in this study. Three observers were invited to contour four structures (bladder, rectum, brain stem and parotid gland) on computed tomography images of the validation cases without any aids. The time taken for contouring was recorded as the manual contouring time. By using an atlas-based auto-segmentation software, three sets of contours were generated for each validation case with different library sizes to produce different degrees of similarity level. The values of the three similarity indices of the auto-segmented contours were calculated. The observers were asked to edit the auto-segmented contours and the editing time was recorded.

The correlation between the editing time and the similarity indices was studied. The amount of time saved was calculated by subtracting the editing time from the manual contouring time. The performances of PBI, DSC and MNAHD were evaluated using Pearson correlation coefficient and receiver operating curve (ROC) analysis.

The PBI showed a positive linear relationship with the amount of contouring time saved. Pearson correlation coefficient ranged from 0·73 to 0·86 for the four structures. The PBI had a stronger correlation than the DSC in bladder and parotid gland, while there was no significant difference between the two indices in rectum and brain stem. The MNAHD had an inferior correlation than the proposed index. For the ROC analysis, the cut-off values for the PBI were 0·549, 0·401 and 0·301 for the three levels of contouring time saved, namely 50, 25 and 0%, respectively. The accuracy of PBI was over 77% and the Youden index was >0·6 for all three levels.

The proposed index showed a stronger relationship to the amount of contouring time saved. It was a simple tool that could be used to evaluate the performance of different segmentation algorithms.

This study aimed to investigate the uncertainty in organ delineation of low-dose computed tomography (CT) images using a high-strength iterative reconstruction (IR) during radiotherapy planning for the treatment of prostate cancer.

Two CT datasets were prepared with different dose levels by adjusting the reconstruction slice thickness. Two observers independently delineated the prostate, seminal vesicles, bladder and rectum on both images without referring to other modality images. The delineated organ volumes were compared between both images. Observer delineation variability was assessed using Dice similarity coefficient (DSC) and mean distance to agreement.

No significant differences regarding the delineated organ volumes were observed between the low- and standard-dose images for all organs. Regarding inter-observer variability, the DSC was relatively high for both images, whereas mean distance to agreement was not significantly different between images (p > 0·05 for all). Intra-observer variability for each observer showed high DSC (>0·8 and >0·9 for seminal vesicles and other organs, respectively) but no significant differences in the mean distance to agreement (p > 0·05 for all).

Our results indicate that low-dose CT images with high-strength IR would be available for organ delineation in the radiotherapy treatment planning for prostate cancer.

Variation in delineation of target volumes/organs at risk (OARs) is well recognised in radiotherapy and may be reduced by several methods including teaching. We evaluated the impact of teaching on contouring variation for thoracic/pelvic stereotactic ablative radiotherapy (SABR) during a virtual contouring workshop.

Target volume/OAR contours produced by workshop participants for three cases were evaluated against reference contours using DICE similarity coefficient (DSC) and line domain error (LDE) metrics. Pre- and post-workshop DSC results were compared using Wilcoxon signed ranks test to determine the impact of teaching during the workshop.

Of 50 workshop participants, paired pre- and post-workshop contours were available for 21 (42%), 20 (40%) and 22 (44%) participants for primary lung cancer, pelvic bone metastasis and pelvic node metastasis cases, respectively. Statistically significant improvements post-workshop in median DSC and LDE results were observed for 6 (50%) and 7 (58%) of 12 structures, respectively, although the magnitude of DSC/LDE improvement was modest in most cases. An increase in median DSC post-workshop ≥0·05 was only observed for GTVbone, IGTVlung and SacralPlex, and reduction in median LDE > 1 mm was only observed for GTVbone, CTVbone and SacralPlex. Post-workshop, median DSC values were >0·7 for 75% of structures. For 92% of the structures, post-workshop contours were considered to be acceptable or within acceptable variation following review by the workshop faculty.

This study has demonstrated that virtual SABR contouring training is feasible and was associated with some improvements in contouring variation for multiple target volumes/OARs.

To investigate the modalities of tumour bed (TB) localisation of target volume delineation [clinically computed tomography (CT), ultrasound (US) compared with surgical clips-guided] and the impact of their differences in delineated TB volumes.

In total, 27 patients who underwent oncoplastic breast conservative surgery with surgical clips insertion (at least three) were included. CT and US imaging for TB localisation were done 3–4 weeks post-operatively in the same treatment position. TB was delineated four times, guided by surgical clips, clinical data, CT (seroma) and US. A plan was done for each TB delineated. The four delineated volumes were compared regarding the volumetric differences, the geographical miss index (GMI) and the overlap index.

Comparing the four modalities, median TB volume was for clinical (60.7), CT (60.8) and US (49.3) cm3, in comparison with 59.7 cm3 for clips, p=0.05. Median of GMI (represented the tissue at risk of recurrence and not had been treated) was for clinical (61.8), CT (45) and US (62.4)%, with significant difference of p=0.02. Median of normal tissue index (normal tissue has been included unnecessarily) was for clinical (59.5), CT (49.6) and US (62.3)%, p=0.17. Overlap index with clips-guided was for clinical (0.36), CT (0.42) and US (0.35) with significance of p=0.04. Median superior/inferior direction was 0.72, −0.03 and −0.2 cm for clinical, CT and US, respectively, with significant value of p=0.02, whereas the anterior–posterior was −0.07, −0.15 and −0.09 cm, p-value=0.45 and the medio–lateral was 0.4, −0.13 and 0.09 cm, p=0.60.

Significant differences in shifts and indices were detected between each of modalities compared with surgical clips. Thus, in the setting of oncoplastic breast surgery, surgical clips should be routinely used for TB localisation. In view of the larger volumes of breast tissue excised and the extensive remodelling that are inherent to oncoplastic procedures, the concept of TB boost irradiation should be re-challenged.

To investigate the use of co-registration of the computerised tomography (CT) planning scan with transperineal ultrasound (TPUS) as an aid to the delineation of the clinical target volume (CTV), and the use of TPUS as a tool for inter- and intra-fractional monitoring in men with bilateral hip prostheses (b-P) undergoing prostate radiotherapy.

We marked the CTV of three patients with and without the co-registered TPUS images. A metal artefact reduction algorithm was utilised. Two patients were treated with intensity-modulated radiotherapy (IMRT) and one with volumetric-modulated arc therapy (VMAT). The inter- and intra-fractional monitoring details were reviewed retrospectively.

Clinician marking with TPUS/CT fusion improved the confidence of prostate CTV delineation leading to a consistent change in volumes across two observers. Inter- and intra-fractional monitoring was possible using TPUS as image guidance, as it is for those patients with non-prosthetic hips.

Using TPUS in the radiotherapy workflow has enabled us to more confidently plan, treat and monitor patients with b-HP. Due to transperineal image acquisition, the ultrasound images are not affected by the presence of hip prostheses, which are outside the field of view.

Magnetic resonance imaging (MRI) is indispensable for treatment planning in prostate radiotherapy (PR). Registration of MRI when compared to planning CT (pCT) is prone to uncertainty and this is rarely reported. In this study, we have compared three different types of registration methods to justify the direct use of MRI in PR.

Thirty patients treated for PR were retrospectively selected for this study and all underwent both CT and MRI. The MR scans were registered to the pCT using markers, focused and unfocussed methods and their registration are REGM, REGF, and REGNF, respectively. Registration comparison is done using the translational differences of three axes from the centre-of-mass values of gross tumour volume (GTV) generated using MRI.

The average difference in all three axes (x, y, z) is (1, 2·5, 2·3 mm) and (1, 3, 2·3 mm) for REGF-REFNF and REGF-REGM, respectively. MR-based GTV Volume is less in comparison to CT-based GTV and it is significantly different (p < 0·001).

Image registration uncertainty is unavoidable for a regular CT–MR workflow. Additional planning target volume margin ranging from 2 to 3mm could be avoided if MR-only workflow is employed. This reduction in the margin is beneficial for small tumours treated with hypofractionation.

This study compared the pattern of radiation induced parotid changes between conventional (ConRT) and intensity modulated radiotherapy (IMRT) in nasopharyngeal carcinoma patients.

56 adult NPC patients treated with IMRT (n=28) and conventional radiotherapy (n=28) were recruited. CT scans were acquired before radiotherapy, at 10th, 20th and 30th fractions, and 3 months after treatment. Parotid gland was delineated in the corresponding CT slices and its mean dose was calculated. The volumetric and geometric changes of the parotid gland at various time intervals were compared against the pre-treatment structure set. The pattern of changes was compared between the two techniques.

The mean parotid dose of IMRT (37.5±9.5 Gy) was significantly lower than ConRT (49.1±7.4 Gy). The parotid gland volume, DICE similarity coefficient and lateral dimension of patient head gradually decreased during the radiotherapy course and partially recovered in 3 months post-treatment. The differences between two groups were not significant until at 3 month after treatment, where IMRT showed significantly better volume recovery.

Similar parotid gland size and location changes were observed during the treatment course in both ConRT and IMRT. However IMRT demonstrated better parotid volume recovery after treatment.

A novel radiotherapy outlining application uses a small number of user-assigned points across orthogonal planes to generate a mesh which is then edited across multiple slices using innovative three-dimensional (3D) sculpting tools. This paper presents the results of a bladder outlining study that compared times and volumes for the new tool with those of a conventional manual outlining tool.

All students undertaking their first University radiotherapy planning module were invited to participate. Following training, they performed a timed outlining of the same male bladder dataset and provided feedback on their preferred method.

Comparison of times from the resulting ten datasets demonstrated that the 3D segmentation tool was significantly faster than conventional software with a mean time of 11·9 minutes compared with 19·2 minutes (p=0·03). The users expressed a preference for the new tool (eight users) over the conventional outlining software (two users).

A minimal point 3D volumetric manual outlining tool utilising orthogonal computed tomography planes demonstrated significant time saving for bladder segmentation compared with axial-based outlining within a group of novice outliners. Future work aims to establish the role of the 3D multi-slice sculpting tools in editing of auto-segmentation derived contour sets.

The aim of the study was to find the dosimetric impact of positron emission tomography (PET)-based gross tumour volume (GTV) delineation over computed tomography (CT)-based GTV delineation for carcinoma oesophagus.

Fifteen patients with carcinoma oesophagus were retrospectively selected. Two sets of GTVs in CT plain images were generated, one with the help of intravenous and oral contrast (GTV CT) and the other with only using PET uptake with the standardised uptake value (simple way of determining the activity in PET) (SUV) > 2.5 (GTV PET). Corresponding PTVs were generated. For all patients, rapid arc plans were generated. Changes in target volumes and critical structure doses were evaluated. The Wilcoxon signed-rank test was used for statistical analysis, and p value < 0.05 was assumed as statistically significant.

Mean reduction in GTV was 5.76 ± 19.35 cc. Mean reduction in PTV 45 Gy was 42.40 ± 76.39 cc. Mean reduction in heart mean dose was 1.53 ± 2.16 Gy. Mean reductions in left lung V20% and V10% were 2.43 ± 4.28 and 3.25 ± 5.09 Gy, respectively. Mean reductions in right lung V20% and V10% were 3.11 ± 4.91 and 2.80 ± 4.51 Gy, respectively. Mean reduction in total lung mean dose was 1.00 ± 1.19 Gy.

PET-based GTV contouring reduces the treatment volume and critical structure doses significantly over CT-based GTV contouring for carcinoma oesophagus.

This study aimed to evaluate and compare simultaneous integrated boost-based volumetric modulated arc therapy (SIB-VMAT) of head-and-neck plans optimised using segmented and non-segmented intermediate-risk target volumes.

CT data of 20 patients with locally advanced laryngeal cancer treated with radical chemoradiation were included retrospectively. Both segmented [planning target volume (PTV) IR!] and non-segmented PTV (PTV IR) volumes were created for the intermediate-risk volume. Correspondingly, two VMAT plans were generated for every CT dataset. Dosimetry parameters obtained from cumulative dose volume histogram and the quality indices such as conformity and homogeneity indices were evaluated for both plans and were statistically analysed.

Maximum dose of PTV IR! was observed to be higher in the non-segmented plans (7281·45 versus 7075·75 cGy) and was statistically significant (p = 0·002). Homogeneity index (HI) of PTV IR! in segmented plans fared better compared to non-segmented plans (0·1 versus 0·12, p = 0·01). All other dosimetry parameters were found to be similar in both plans.

This study shows that using segmented volumes for planning will lead to more homogenous plans with regard to intermediate- and low-risk volumes, especially under controlled settings.

Head and neck cancer (HNC) accounts for 5% of all new cancer cases and most were locally advanced. Positron emission tomography/computed tomography (PET/CT) in radiotherapy practice in locally advanced head and neck squamous cell carcinoma (HNSCC) is being used in staging and proper contouring. Proper staging is essential for accurate treatment decision.

This is a prospective phase II study conducted as a single institute centre to evaluate the role of PET/CT-treatment in staging, contouring and response evaluation of 30 patients with locally advanced HNSCC in contrast to CT scan. Our cases did not undergo radical surgery for the primary tumour, and biopsy was taken with PET/CT post-treatment to evaluate response.

Median age of patients was 49·4 years (minimum age of 32 years and maximum of 68 years). Males were predominant 22 (73·3%). Nasopharynx was the predominant site 16/30 (53·3%). PET/CT changed the overall staging in 40% of the patients (upstaged in 36·7% and downstaged in 3·3%). Gross tumour volume (GTV) of PET/CT was smaller in 23 patients (76·7%) and larger in 5 (16·7%) than the GTV of conventional CT, whereas GTV of lymph nodes of PET/CT was larger in 20 patients (67·7%) and smaller in 4 (13·3%). PET/CT study detected bone metastasis in two nasopharyngeal carcinoma patients and two cases of 2nd primary tumours which were not detected using conventional CT. The Cox-regression model showed that the median standardised uptake volume (SUV) of the initial tumour had been a dependent predictor of death in patients with HNSCC (p-value = 0·033) where the risk of death was 0·725 times among patients with high SUV of the initial tumour. Consequently, the size of GTV of the tumour was significant in the prediction of death (p = 0·018).

18F-FDG-PET/CT is useful for staging, radiotherapy delineation as well as aiding proper decision making, in addition to assessment of treatment response in HNSCC patients.

Paratesticular sarcoma are extremely rare malignant tumours. Unlike other sites, they tend to be lower grade and have higher propensity of lymphatic spread. They tend to fail locally and occasionally in the regional lymph nodes. In the absence of target volume delineation guidelines and technical illustration of conformal planning, we have made an attempt to illustrate conformal planning methodology and define target volume based on current evidence in a case of paratesticular sarcoma.

We are presenting a case of 62-year-old male who presented with 15-cm scrotal swelling and underwent high inguinal orchidectomy with ligation of spermatic cord. Histopathology presented a well-differentiated leiomyosarcoma of epididymis. Post-operative radiotherapy target volume included the tumour bed, ipsilateral inguinal nodes and lower pelvic nodes as the clinical target volume.

Adjuvant radiotherapy using advanced delivery technique such as volumetric arc technique can provide good dose distribution with good sparing of organs at risk. The downside of conformal radiation delivery is that it is a resource-intensive and has no established target volume delineation guidelines.

The aim of this systematic review was to synthesise and summarise evidence surrounding the clinical use of fluoro-2-deoxy-d-glucose positron emission tomography/computed tomography (FDG-PET/CT) and magnetic resonance imaging (MRI) for target volume delineation (TVD) in rectal cancer radiotherapy planning.

PubMed, EMBASE, Cochrane library, CINAHL, Web of Science and Scopus databases and other sources were systematically queried using keywords and relevant synonyms. Eligible full-text studies were assessed for methodological quality using the QUADAS-2 tool.

Eight of the 1448 studies identified met the inclusion criteria. Findings showed that MRI significantly delineate larger tumour volumes (TVs) than FDG-PET/CT while diffusion-weighted magnetic resonance imaging (DW-MRI) defined smaller gross tumour volumes (GTVs) compared to T2 weighted-Magnetic Resonance Image. CT-based GTVs were found to be larger compared to FDG-PET/CT. FDG-PET/CT also identified new lesions in 15–17% patients and TVs extending outside the routinely used clinical standard CT TV in 29–83% patients. Between observers, delineated volumes were similar and consistent between MRI sequences, whereas interobserver agreement was significantly improved with FDG-PET/CT than CT.

FDG-PET/CT and DW-MRI appear to delineate smaller rectal TVs and show improved interobserver variability. Overall, this study provides valuable insights into the amount of attention in the research literature that has been paid to imaging for TVD in rectal cancer.