ABSTRACT
OBJECTIVE: To test the ability of high-performance machine learning (ML) models employing clinical, radiological, and radiomic variables to improve non-invasive prediction of the pathological status of prostate cancer (PCa) in a large, single-institution cohort. METHODS: Patients who underwent multiparametric MRI and prostatectomy in our institution in 2015-2018 were considered; a total of 949 patients were included. Gradient-boosted decision tree models were separately trained using clinical features alone and in combination with radiological reporting and/or prostate radiomic features to predict pathological T, pathological N, ISUP score, and their change from preclinical assessment. Model behavior was analyzed in terms of performance, feature importance, Shapley additive explanation (SHAP) values, and mean absolute error (MAE). The best model was compared against a naïve model mimicking clinical workflow. RESULTS: The model including all variables was the best performing (AUC values ranging from 0.73 to 0.96 for the six endpoints). Radiomic features brought a small yet measurable boost in performance, with the SHAP values indicating that their contribution can be critical to successful prediction of endpoints for individual patients. MAEs were lower for low-risk patients, suggesting that the models find them easier to classify. The best model outperformed (p ≤ 0.0001) clinical baseline, resulting in significantly fewer false negative predictions and overall was less prone to under-staging. CONCLUSIONS: Our results highlight the potential benefit of integrative ML models for pathological status prediction in PCa. Additional studies regarding clinical integration of such models can provide valuable information for personalizing therapy offering a tool to improve non-invasive prediction of pathological status. CLINICAL RELEVANCE STATEMENT: The best machine learning model was less prone to under-staging of the disease. The improved accuracy of our pathological prediction models could constitute an asset to the clinical workflow by providing clinicians with accurate pathological predictions prior to treatment. KEY POINTS: ⢠Currently, the most common strategies for pre-surgical stratification of prostate cancer (PCa) patients have shown to have suboptimal performances. ⢠The addition of radiological features to the clinical features gave a considerable boost in model performance. Our best model outperforms the naïve model, avoiding under-staging and resulting in a critical advantage in the clinic. â¢Machine learning models incorporating clinical, radiological, and radiomics features significantly improved accuracy of pathological prediction in prostate cancer, possibly constituting an asset to the clinical workflow.
ABSTRACT
The work investigates the implementation of personalized radiotherapy boluses by means of additive manufacturing technologies. Boluses materials that are currently used need an excessive amount of human intervention which leads to reduced repeatability in terms of dosimetry. Additive manufacturing can solve this problem by eliminating the human factor in the process of fabrication. Planar boluses with fixed geometry and personalized boluses printed starting from a computed tomography scan of a radiotherapy phantom were produced. First, a dosimetric characterization study on planar bolus designs to quantify the effects of print parameters such as infill density and geometry on the radiation beam was made. Secondly, a volumetric quantification of air gap between the bolus and the skin of the patient as well as dosimetric analyses were performed. The optimization process according to the obtained dosimetric and airgap results allowed us to find a combination of parameters to have the 3D-printed bolus performing similarly to that in conventional use. These preliminary results confirm those in the relevant literature, with 3D-printed boluses showing a dosimetric performance similar to conventional boluses with the additional advantage of being perfectly conformed to the patient geometry.
Subject(s)
Phantoms, Imaging , Printing, Three-Dimensional , Radiometry , Humans , Radiotherapy Dosage , Tomography, X-Ray Computed , Air , Radiotherapy/methods , Radiotherapy/instrumentationABSTRACT
PURPOSE: Hippocampal sparing whole-brain radiotherapy (HS-WBRT) showed significantly lower long-term side effects compared to standard WBRT. Aim of this study is to describe a HS-WBRT real-world monoinstitutional experience within a retrospective cohort. METHODS: Patients who completed HS-WBRT course, with Karnofsky Performance Status ⩾ 60 and radiological diagnosis of brain metastases (BMs) were enrolled. Treatment was performed using helical Tomotherapy scheduled in 30 Gy in 10 or 12 fractions or 25 Gy in 10 fractions. Oncological outcomes were clinically and radiologically assessed every three months. Toxicity was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events 4.3. RESULTS: One hundred and nineteen patients from 2016 to 2020 met inclusion criteria; after a median follow-up of 18 months, 29 patients were alive; 6- and 12-months overall survival rates were 66% and 41%, respectively. HS-WBRT response was assessed for 72 patients. Median time to any progression and intracranial failure (IF) was 4.5 and 13.7 months, respectively. The 6- and 12-month IF rates were 85% and 57%. Among 40 patients (34%) who experienced IF, 17 (42%) were oligometastatic, 23 (58%) polymetastatic and 15/40 developed IF within the hippocampi avoidance zone. No grade (G) ⩾ 2 acute toxicities were reported and one G2 (dizziness) late toxicity was described. CONCLUSIONS: HS-WBRT is well tolerated, and despite the hippocampal sparing region, the oncological control is satisfying. Further investigation is warranted to find patients who could most benefit from a HS-WBRT approach.
Subject(s)
Brain Neoplasms , Radiotherapy, Intensity-Modulated , Humans , Retrospective Studies , Feasibility Studies , Radiotherapy Planning, Computer-Assisted , Cranial Irradiation/adverse effects , Brain Neoplasms/pathology , Radiotherapy, Intensity-Modulated/adverse effects , Hippocampus/pathologyABSTRACT
AIM: Radiation-induced oral mucositis (RIOM) is the most frequent side effect in head and neck cancer (HNC) patients treated with curative radiotherapy (RT). A standardized strategy for preventing and treating RIOM has not been defined. Aim of this study was to perform a real-life survey on RIOM management among Italian RT centers. METHODS: A 40-question survey was administered to 25 radiation oncologists working in 25 different RT centers across Italy. RESULTS: A total of 1554 HNC patients have been treated in the participating centers in 2021, the majority (median across the centers 91%) with curative intent. Median treatment time was 41 days, with a mean percentage of interruption due to toxicity of 14.5%. Eighty percent of responders provide written oral cavity hygiene recommendations. Regarding RIOM prevention, sodium bicarbonate mouthwashes, oral mucosa barrier agents, and hyaluronic acid-based mouthwashes were the most frequent topic agents used. Regarding RIOM treatment, 14 (56%) centers relied on literature evidence, while internal guidelines were available in 13 centers (44%). Grade (G)1 mucositis is mostly treated with sodium bicarbonate mouthwashes, oral mucosa barrier agents, and steroids, while hyaluronic acid-based agents, local anesthetics, and benzydamine were the most used in mucositis G2/G3. Steroids, painkillers, and anti-inflammatory drugs were the most frequent systemic agents used independently from the RIOM severity. CONCLUSION: Great variety of strategies exist among Italian centers in RIOM management for HNC patients. Whether different strategies could impact patients' compliance and overall treatment time of the radiation course is still unclear and needs further investigation.
Subject(s)
Head and Neck Neoplasms , Mucositis , Radiation Injuries , Radiation Oncology , Stomatitis , Humans , Mucositis/drug therapy , Mouthwashes/therapeutic use , Sodium Bicarbonate/therapeutic use , Hyaluronic Acid/therapeutic use , Stomatitis/etiology , Stomatitis/prevention & control , Radiation Injuries/etiology , Radiation Injuries/prevention & control , Head and Neck Neoplasms/radiotherapy , Head and Neck Neoplasms/drug therapy , SteroidsABSTRACT
This study quantified the incidental dose to the first axillary level (L1) in locoregional treatment plan for breast cancer. Eighteen radiotherapy centres contoured L1-L4 on three different patients (P1,2,3), created the L2-L4 planning target volume (single centre planning target volume, SC-PTV) and elaborated a locoregional treatment plan. The L2-L4 gold standard clinical target volume (CTV) along with the gold standard L1 contour (GS-L1) were created by an expert consensus. The SC-PTV was then replaced by the GS-PTV and the incidental dose to GS-L1 was measured. Dosimetric data were analysed with Kruskal-Wallis test. Plans were intensity modulated radiotherapy (IMRT)-based. P3 with 90° arm setup had statistically significant higher L1 dose across the board than P1 and P2, with the mean dose (Dmean) reaching clinical significance. Dmean of P1 and P2 was consistent with the literature (77.4% and 74.7%, respectively). The incidental dose depended mostly on L1 proportion included in the breast fields, underlining the importance of the setup, even in case of IMRT.
Subject(s)
Breast Neoplasms , Radiotherapy, Intensity-Modulated , Humans , Female , Breast Neoplasms/radiotherapy , Radiotherapy Planning, Computer-Assisted , Radiotherapy Dosage , Observer Variation , BreastABSTRACT
AIMS: To assess whether CT-based radiomics and blood-derived biomarkers could improve the prediction of overall survival (OS) and locoregional progression-free survival (LRPFS) in patients with oropharyngeal cancer (OPC) treated with curative-intent RT. METHODS: Consecutive OPC patients with primary tumors treated between 2005 and 2021 were included. Analyzed clinical variables included gender, age, smoking history, staging, subsite, HPV status, and blood parameters (baseline hemoglobin levels, neutrophils, monocytes, and platelets, and derived measurements). Radiomic features were extracted from the gross tumor volumes (GTVs) of the primary tumor using pyradiomics. Outcomes of interest were LRPFS and OS. Following feature selection, a radiomic score (RS) was calculated for each patient. Significant variables, along with age and gender, were included in multivariable analysis, and models were retained if statistically significant. The models' performance was compared by the C-index. RESULTS: One hundred and five patients, predominately male (71%), were included in the analysis. The median age was 59 (IQR: 52-66) years, and stage IVA was the most represented (70%). HPV status was positive in 63 patients, negative in 7, and missing in 35 patients. The median OS follow-up was 6.3 (IQR: 5.5-7.9) years. A statistically significant association between low Hb levels and poorer LRPFS in the HPV-positive subgroup (p = 0.038) was identified. The calculation of the RS successfully stratified patients according to both OS (log-rank p < 0.0001) and LRPFS (log-rank p = 0.0002). The C-index of the clinical and radiomic model resulted in 0.82 [CI: 0.80-0.84] for OS and 0.77 [CI: 0.75-0.79] for LRPFS. CONCLUSIONS: Our results show that radiomics could provide clinically significant informative content in this scenario. The best performances were obtained by combining clinical and quantitative imaging variables, thus suggesting the potential of integrative modeling for outcome predictions in this setting of patients.
ABSTRACT
BACKGROUND: Radiomics represents an emerging field of precision-medicine. Its application in head and neck is still at the beginning. METHODS: Retrospective study about magnetic resonance imaging (MRI) based radiomics in oral tongue squamous cell carcinoma (OTSCC) surgically treated (2010-2019; 79 patients). All preoperative MRIs include different sequences (T1, T2, DWI, ADC). Tumor volume was manually segmented and exported to radiomic-software, to perform feature extraction. Statistically significant variables were included in multivariable analysis and related to survival endpoints. Predictive models were elaborated (clinical, radiomic, clinical-radiomic models) and compared using C-index. RESULTS: In almost all clinical-radiomic models radiomic-score maintained statistical significance. In all cases C-index was higher in clinical-radiomic models than in clinical ones. ADC provided the best fit to the models (C-index 0.98, 0.86, 0.84 in loco-regional recurrence, cause-specific mortality, overall survival, respectively). CONCLUSION: MRI-based radiomics in OTSCC represents a promising noninvasive method of precision medicine, improving prognosis prediction before surgery.
Subject(s)
Carcinoma, Squamous Cell , Head and Neck Neoplasms , Tongue Neoplasms , Humans , Retrospective Studies , Tongue Neoplasms/diagnostic imaging , Carcinoma, Squamous Cell/diagnostic imaging , Carcinoma, Squamous Cell/therapy , Prognosis , Magnetic Resonance Imaging/methods , Squamous Cell Carcinoma of Head and NeckABSTRACT
BACKGROUND: Contouring of anatomical regions is a crucial step in the medical workflow and is both time-consuming and prone to intra- and inter-observer variability. This study compares different strategies for automatic segmentation of the prostate in T2-weighted MRIs. METHODS: This study included 100 patients diagnosed with prostate adenocarcinoma who had undergone multi-parametric MRI and prostatectomy. From the T2-weighted MR images, ground truth segmentation masks were established by consensus from two expert radiologists. The prostate was then automatically contoured with six different methods: (1) a multi-atlas algorithm, (2) a proprietary algorithm in the Syngo.Via medical imaging software, and four deep learning models: (3) a V-net trained from scratch, (4) a pre-trained 2D U-net, (5) a GAN extension of the 2D U-net, and (6) a segmentation-adapted EfficientDet architecture. The resulting segmentations were compared and scored against the ground truth masks with one 70/30 and one 50/50 train/test data split. We also analyzed the association between segmentation performance and clinical variables. RESULTS: The best performing method was the adapted EfficientDet (model 6), achieving a mean Dice coefficient of 0.914, a mean absolute volume difference of 5.9%, a mean surface distance (MSD) of 1.93 pixels, and a mean 95th percentile Hausdorff distance of 3.77 pixels. The deep learning models were less prone to serious errors (0.854 minimum Dice and 4.02 maximum MSD), and no significant relationship was found between segmentation performance and clinical variables. CONCLUSIONS: Deep learning-based segmentation techniques can consistently achieve Dice coefficients of 0.9 or above with as few as 50 training patients, regardless of architectural archetype. The atlas-based and Syngo.via methods found in commercial clinical software performed significantly worse (0.855[Formula: see text]0.887 Dice).
Subject(s)
Prostate , Prostatic Neoplasms , Male , Humans , Prostate/diagnostic imaging , Image Processing, Computer-Assisted/methods , Algorithms , Prostatic Neoplasms/diagnostic imaging , Magnetic Resonance Imaging/methodsABSTRACT
BACKGROUND AND PURPOSE: Radiomics enables the mining of quantitative features from medical images. The influence of the radiomic feature extraction software on the final performance of models is still a poorly understood topic. This study aimed to investigate the ability of radiomic features extracted by two different radiomic platforms to predict clinical outcomes in patients treated with radiosurgery for brain metastases from non-small cell lung cancer. We developed models integrating pre-treatment magnetic resonance imaging (MRI)-derived radiomic features and clinical data. MATERIALS AND METHODS: Pre-radiotherapy gadolinium enhanced axial T1-weighted MRI scans were used. MRI images were re-sampled, intensity-shifted, and histogram-matched before radiomic extraction by means of two different platforms (PyRadiomics and SOPHiA Radiomics). We adopted LASSO Cox regression models for multivariable analyses by creating radiomic, clinical, and combined models using three survival clinical endpoints (local control, distant progression, and overall survival). The statistical analysis was repeated 50 times with different random seeds and the median concordance index was used as performance metric of the models. RESULTS: We analysed 276 metastases from 148 patients. The use of the two platforms resulted in differences in both the quality and the number of extractable features. That led to mismatches in terms of end-to-end performance, statistical significance of radiomic scores, and clinical covariates found significant in combined models. CONCLUSION: This study shed new light on how extracting radiomic features from the same images using two different platforms could yield several discrepancies. That may lead to acute consequences on drawing conclusions, comparing results across the literature, and translating radiomics into clinical practice.
Subject(s)
Brain Neoplasms , Carcinoma, Non-Small-Cell Lung , Lung Neoplasms , Radiosurgery , Humans , Carcinoma, Non-Small-Cell Lung/diagnostic imaging , Carcinoma, Non-Small-Cell Lung/radiotherapy , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/radiotherapy , Lung Neoplasms/etiology , Radiosurgery/methods , Brain Neoplasms/diagnostic imaging , Brain Neoplasms/radiotherapy , Magnetic Resonance Imaging/methods , Retrospective StudiesABSTRACT
Radiotherapy represents a highly targeted and efficient treatment choice in many cancer types, both with curative and palliative intents. Nevertheless, radioresistance, consisting in the adaptive response of the tumor to radiation-induced damage, represents a major clinical problem. A growing body of the literature suggests that mechanisms related to mitochondrial changes and metabolic remodeling might play a major role in radioresistance development. In this work, the main contributors to the acquired cellular radioresistance and their relation with mitochondrial changes in terms of reactive oxygen species, hypoxia, and epigenetic alterations have been discussed. We focused on recent findings pointing to a major role of mitochondria in response to radiotherapy, along with their implication in the mechanisms underlying radioresistance and radiosensitivity, and briefly summarized some of the recently proposed mitochondria-targeting strategies to overcome the radioresistant phenotype in cancer.
Subject(s)
Neoplasms , Cell Line, Tumor , Humans , Hypoxia/metabolism , Mitochondria/metabolism , Neoplasms/genetics , Neoplasms/metabolism , Neoplasms/radiotherapy , Radiation Tolerance/genetics , Reactive Oxygen Species/metabolismABSTRACT
OBJECTIVE: Deploying an automatic segmentation model in practice should require rigorous quality assurance (QA) and continuous monitoring of the model's use and performance, particularly in high-stakes scenarios such as healthcare. Currently, however, tools to assist with QA for such models are not available to AI researchers. In this work, we build a deep learning model that estimates the quality of automatically generated contours. METHODS: The model was trained to predict the segmentation quality by outputting an estimate of the Dice similarity coefficient given an image contour pair as input. Our dataset contained 60 axial T2-weighted MRI images of prostates with ground truth segmentations along with 80 automatically generated segmentation masks. The model we used was a 3D version of the EfficientDet architecture with a custom regression head. For validation, we used a fivefold cross-validation. To counteract the limitation of the small dataset, we used an extensive data augmentation scheme capable of producing virtually infinite training samples from a single ground truth label mask. In addition, we compared the results against a baseline model that only uses clinical variables for its predictions. RESULTS: Our model achieved a mean absolute error of 0.020 ± 0.026 (2.2% mean percentage error) in estimating the Dice score, with a rank correlation of 0.42. Furthermore, the model managed to correctly identify incorrect segmentations (defined in terms of acceptable/unacceptable) 99.6% of the time. CONCLUSION: We believe that the trained model can be used alongside automatic segmentation tools to ensure quality and thus allow intervention to prevent undesired segmentation behavior.
ABSTRACT
Aim: The purpose of this study is to collect available evidence on the feasibility and efficacy of stereotactic arrhythmia radio ablation (STAR), including both photon radiotherapy (XRT) and particle beam therapy (PBT), in the treatment of atrial fibrillation (AF), and to provide cardiologists and radiation oncologists with a practical overview on this topic. Methods: Three hundred and thirty-five articles were identified up to November 2021 according to preferred reporting items for systematic reviews and meta-analyses criteria; preclinical and clinical studies were included without data restrictions or language limitations. Selected works were analyzed for comparing target selection, treatment plan details, and the accelerator employed, addressing workup modalities, acute and long-term side-effects, and efficacy, defined either by the presence of scar or by the absence of AF recurrence. Results: Twenty-one works published between 2010 and 2021 were included. Seventeen studies concerned XRT, three PBT, and one involved both. Nine studies (1 in silico and 8 in vivo; doses ranging from 15 to 40 Gy) comprised a total of 59 animals, 12 (8 in silico, 4 in vivo; doses ranging from 16 to 50 Gy) focused on humans, with 9 patients undergoing STAR: average follow-up duration was 5 and 6 months, respectively. Data analysis supported efficacy of the treatment in the preclinical setting, whereas in the context of clinical studies the main favorable finding consisted in the detection of electrical scar in 4/4 patients undergoing specific evaluation; the minimum dose for efficacy was 25 Gy in both humans and animals. No acute complication was recorded; severe side-effects related to the long-term were observed only for very high STAR doses in 2 animals. Significant variability was evidenced among studies in the definition of target volume and doses, and in the management of respiratory and cardiac target motion. Conclusion: STAR is an innovative non-invasive procedure already applied for experimental treatment of ventricular arrhythmias. Particular attention must be paid to safety, rather than efficacy of STAR, given the benign nature of AF. Uncertainties persist, mainly regarding the definition of the treatment plan and the role of the target motion. In this setting, more information about the toxicity profile of this new approach is compulsory before applying STAR to AF in clinical practice.
ABSTRACT
BACKGROUND AND PURPOSE: Diversity, Equity and Inclusion (DEI) in the medical workforce is linked to improved patient care and innovation, as well as employee retention and engagement. The European Society for Radiotherapy and Oncology launched a survey to provide a benchmark of DEI and engagement among radiation oncology (RO) professionals in Europe. METHODS: An anonymous survey was disseminated among RO professionals in Europe. The survey collected demographics and professional information, and participants were asked if they felt they belonged to a minority group. A DEI and workforce engagement questionnaire by Person et al. evaluated 8 inclusion factors. A favourable score was calculated by adding the percentage of "strongly agreed" or "agreed" answers. RESULTS: A total of 812 complete responses were received from 35 European countries. 21% of respondents felt they belonged to a minority group, mostly based on race/ethnicity (5.9%), nationality (4.8%) and age (4.3%). Compared to benchmark data from the United States, scores were lower for most inclusion factors, and to a greater extent for minority groups. The overall favourable score was 58% for those belonging to a minority group, significantly lower than for other respondents (71%, p < 0.001). Those belonging to a minority group because of their gender or age had the lowest overall favourable score (47% and 51% respectively). CONCLUSIONS: Our work indicates that actions to improve DEI and workforce engagement among RO professionals in Europe are urgently needed, in particular among minority groups. This would potentially improve employee wellbeing and retention, promoting high quality care and innovation.
Subject(s)
Radiation Oncology , Benchmarking , Europe , Humans , Minority Groups , United States , WorkforceABSTRACT
After primary treatment for prostate cancer with either radical prostatectomy or radiotherapy, a significant proportion of patients are at risk of developing metastases. In recent years, a deeper understanding of the underlying biology together with improved imaging techniques and the advent of new therapeutic options including metastases-directed therapies and new drugs have revolutionized the management of low-burden metastatic disease, also known as oligometastatic state. The purpose of this narrative review is to report the recent developments in the management of hormone-sensitive oligometastatic prostate cancer patients.
ABSTRACT
AIM: To quantify the dosimetric impact of contouring variability of axillary lymph nodes (L2, L3, L4) in breast cancer (BC) locoregional radiotherapy (RT). MATERIALS AND METHODS: 18 RT centres were asked to plan a locoregional treatment on their own planning target volume (single centre, SC-PTV) which was created by applying their institutional margins to the clinical target volume of the axillary nodes of three BC patients (P1, P2, P3) previously delineated (SC-CTV). The gold standard CTVs (GS-CTVs) of P1, P2 and P3 were developed by BC experts' consensus and validated with STAPLE algorithm. For each participating centre, the GS-PTV of each patient was created by applying the same margins as those used for the SC-CTV to SC-PTV expansion and replaced the SC-PTV in the treatment plan. Datasets were imported into MIM v6.1.7 [MIM Software Inc.], where dose-volume histograms (DVHs) were extracted and differences were analysed. RESULTS: 17/18 centres used intensity-modulated RT (IMRT). The CTV to PTV margins ranged from 0 to 10 mm (median 5 mm). No correlation was observed between GS-CTV coverage by 95% isodose and GS-PTV margins width. Doses delivered to 98% (D98) and 95% (D95) of GS-CTVs were significantly lower than those delivered to the SC-CTVs. No significant difference between SC-CTV and GS-CTV was observed in maximum dose (D2), always under 110%. Mean dose ≥99% of the SC-CTVs and GS-CTVs was satisfied in 84% and 50%, respectively. In less than one half of plans, GS-CTV V95% was above 90%. Breaking down the GS-CTV into the three nodal levels (L2, L3 and L4), L4 had the lowest probability to be covered by the 95% isodose. CONCLUSIONS: Overall, GS-CTV resulted worse coverage, especially for L4. IMRT was largely used and CTV-to-PTV margins did not compensate for contouring issues. The results highlighted the need for delineation training and standardization.
Subject(s)
Breast Neoplasms , Radiotherapy, Intensity-Modulated , Breast Neoplasms/radiotherapy , Female , Humans , Lymph Nodes , Radiometry/methods , Radiotherapy Dosage , Radiotherapy Planning, Computer-Assisted/methods , Radiotherapy, Intensity-Modulated/methodsABSTRACT
Background: No evidence supports the choice of specific imaging filtering methodologies in radiomics. As the volume of the primary tumor is a well-recognized prognosticator, our purpose is to assess how filtering may impact the feature/volume dependency in computed tomography (CT) images of non-small cell lung cancer (NSCLC), and if such impact translates into differences in the performance of survival modeling. The role of lesion volume in model performances was also considered and discussed. Methods: Four-hundred seventeen CT images NSCLC patients were retrieved from the NSCLC-Radiomics public repository. Pre-processing and features extraction were implemented using Pyradiomics v3.0.1. Features showing high correlation with volume across original and filtered images were excluded. Cox proportional hazards (PH) with least absolute shrinkage and selection operator (LASSO) regularization and CatBoost models were built with and without volume, and their concordance (C-) indices were compared using Wilcoxon signed-ranked test. The Mann Whitney U test was used to assess model performances after stratification into two groups based on low- and high-volume lesions. Results: Radiomic models significantly outperformed models built on only clinical variables and volume. However, the exclusion/inclusion of volume did not generally alter the performances of radiomic models. Overall, performances were not substantially affected by the choice of either imaging filter (overall C-index 0.539-0.590 for Cox PH and 0.589-0.612 for CatBoost). The separation of patients with high-volume lesions resulted in significantly better performances in 2/10 and 7/10 cases for Cox PH and CatBoost models, respectively. Both low- and high-volume models performed significantly better with the inclusion of radiomic features (P<0.0001), but the improvement was largest in the high-volume group (+10.2% against +8.7% improvement for CatBoost models and +10.0% against +5.4% in Cox PH models). Conclusions: Radiomic features complement well-known prognostic factors such as volume, but their volume-dependency is high and should be managed with vigilance. The informative content of radiomic features may be diminished in small lesion volumes, which could limit the applicability of radiomics in early-stage NSCLC, where tumors tend to be small. Our results also suggest an advantage of CatBoost models over the Cox PH models.
ABSTRACT
The advent of technologies allowing the global analysis of biological phenomena, referred to as "omics" (genomics, epigenomics, proteomics, metabolomics, microbiomics, radiomics, and radiogenomics), has revolutionized the study of human diseases and traced the path for quantitative personalized medicine. The newly inaugurated Master of Science Program in Biomedical Omics of the University of Milan, Italy, aims at addressing the unmet need to create professionals with a broad understanding of omics disciplines. The course is structured over 2 years and admits students with a bachelor's degree in biotechnology, biology, chemistry, or pharmaceutical sciences. All teaching activities are fully held in English. A total of nine students enrolled in the first academic year and attended the courses of radiomics, genomics and epigenomics, proteomics, and high-throughput screenings, and their feedback was evaluated by means of an online questionnaire. Faculty with different backgrounds were recruited according to the subject. Due to restrictions imposed by the coronavirus disease 2019 (COVID-19) pandemic, laboratory activities were temporarily suspended, while lectures, journal clubs, and examinations were mainly held online. After the end of the first semester, despite the difficulties brought on by the COVID-19 pandemic, the course overall met the expectations of the students, specifically regarding teaching effectiveness, interpersonal interactions with the lecturers, and courses organization. Future efforts will be undertaken to better calibrate the overall workload of the course and to implement the most relevant suggestions from the students together with omics science evolution in order to guarantee state-of-the-art omics teaching and to prepare future omics specialists.
Subject(s)
Biomedical Research/education , COVID-19/genetics , Pandemics/prevention & control , SARS-CoV-2/genetics , COVID-19/virology , Epigenomics/education , Genomics/education , Humans , Metabolomics/education , Proteomics/education , SARS-CoV-2/pathogenicityABSTRACT
Lombardy has represented the Italian and European epicenter of the coronavirus disease 2019 (COVID-19) pandemic. Although most clinical efforts within hospitals were diverted towards the care of virally infected patients, therapies for patients with cancer, including radiotherapy (RT), have continued. During both the first and second pandemic waves, several national and regional organizations provided Italian and Lombardian RT departments with detailed guidelines aimed at ensuring safe treatments during the pandemic. The spread of infection among patients and personnel was limited by adopting strict measures, including triage procedures, interpersonal distance, and adequate implementation of personal protective equipment (PPE). Screening procedures addressed to both the healthcare workforce and patients, such as periodic nasopharyngeal swabs, have allowed the early identification of asymptomatic or pauci-symptomatic COVID-19 cases, thus reducing the spread of the infection. Prevention of infection was deemed of paramount importance to protect both patients and personnel and to ensure the availability of a minimum number of staff members to maintain clinical activity. The choice of treating COVID-19-positive patients has represented a matter of debate, and the risk of oncologic progression has been weighted against the risk of infection of personnel and other patients. Such risk was minimized by creating dedicated paths, reserving time slots, applying intensified cleaning procedures, and supplying personnel and staff with appropriate PPE. Remote working of research staff, medical physicists, and, in some cases, radiation oncologists has prevented overcrowding of shared spaces, reducing infection spread.
Subject(s)
COVID-19 , Neoplasms , Radiation Oncology , COVID-19/epidemiology , Humans , Italy/epidemiology , Neoplasms/epidemiology , Neoplasms/radiotherapy , Pandemics/prevention & control , Personal Protective Equipment , SARS-CoV-2ABSTRACT
AIM: The primary aim of this study is to provide preliminary indications for safe constraints of rectum and bladder in patients re-irradiated with stereotactic body RT (SBRT). METHODS: Data from patients treated for prostate cancer (PCa) and intraprostatic relapse, from 1998 to 2016, were retrospectively collected. First RT course was delivered with 3D conformal RT techniques, SBRT or volumetric modulated arc therapy (VMAT). All patients underwent re-irradiation with SBRT with heavy hypofractionated schedules. Cumulative dose-volume values to organs at risk (OARs) were computed and possible correlation with developed toxicities was investigated. RESULTS: Twenty-six patients were included. Median age at re-irradiation was 75 years, mean interval between the two RT courses was 5.6 years and the median follow-up was 47.7 months (13.4-114.3 months). After re-irradiation, acute and late G ≥ 2 GU toxicity events were reported in 3 (12%) and 10 (38%) patients, respectively, while late G ≥ 2 GI events were reported in 4 (15%) patients. No acute G ≥ 2 GI side effects were registered. Patients receiving an equivalent uniform dose of the two RT treatments < 131 Gy appeared to be at higher risk of progression (4-yr b-PFS: 19% vs 33%, p = 0.145). Cumulative re-irradiation constraints that appear to be safe are D30% < 57.9 Gy for bladder and D30% < 66.0 Gy, D60% < 38.0 Gy and V122.1 Gy < 5% for rectum. CONCLUSION: Preliminary re-irradiation constraints for bladder and rectum have been reported. Our preliminary investigation may serve to clear some grey areas of PCa re-irradiation.
Subject(s)
Prostatic Neoplasms , Radiosurgery , Re-Irradiation , Male , Humans , Child , Re-Irradiation/adverse effects , Re-Irradiation/methods , Radiosurgery/adverse effects , Radiosurgery/methods , Retrospective Studies , Neoplasm Recurrence, Local/radiotherapy , Neoplasm Recurrence, Local/etiology , Prostatic Neoplasms/radiotherapyABSTRACT
BACKGROUND AND PURPOSE: Machine learning (ML) is emerging as a feasible approach to optimize patients' care path in Radiation Oncology. Applications include autosegmentation, treatment planning optimization, and prediction of oncological and toxicity outcomes. The purpose of this clinically oriented systematic review is to illustrate the potential and limitations of the most commonly used ML models in solving everyday clinical issues in head and neck cancer (HNC) radiotherapy (RT). MATERIALS AND METHODS: Electronic databases were screened up to May 2021. Studies dealing with ML and radiomics were considered eligible. The quality of the included studies was rated by an adapted version of the qualitative checklist originally developed by Luo et al. All statistical analyses were performed using R version 3.6.1. RESULTS: Forty-eight studies (21 on autosegmentation, four on treatment planning, 12 on oncological outcome prediction, 10 on toxicity prediction, and one on determinants of postoperative RT) were included in the analysis. The most common imaging modality was computed tomography (CT) (40%) followed by magnetic resonance (MR) (10%). Quantitative image features were considered in nine studies (19%). No significant differences were identified in global and methodological scores when works were stratified per their task (i.e., autosegmentation). DISCUSSION AND CONCLUSION: The range of possible applications of ML in the field of HN Radiation Oncology is wide, albeit this area of research is relatively young. Overall, if not safe yet, ML is most probably a bet worth making.