Search

Simple Summary: Studies comparing performances of machine learning (ML) methods in building predictive models of toxicity in RT are rare. Thanks to the availability of a large cohort (n = 1314) of breast cancer patients homogeneously treated with tangential fields, different ML approaches could be compared. This work shows how more complex models typically achieve higher performances. At the same time, for this test case, the importance is given mainly by a few variables, and toxicity can be predicted by simpler models with similar performances. The availability of more individually characterizing features (here partially missing) is expected to have a likely much higher impact than the choice of the best-performing ML/DL approach. Purpose. Different ML models were compared to predict toxicity in RT on a large cohort (n = 1314). Methods. The endpoint was RTOG G2/G3 acute toxicity, resulting in 204/1314 patients with the event. The dataset, including 25 clinical, anatomical, and dosimetric features, was split into 984 for training and 330 for internal tests. The dataset was standardized; features with a high p-value at univariate LR and with Spearman ρ > 0.8 were excluded; synthesized data of the minority were generated to compensate for class imbalance. Twelve ML methods were considered. Model optimization and sequential backward selection were run to choose the best models with a parsimonious feature number. Finally, feature importance was derived for every model. Results. The model's performance was compared on a training–test dataset over different metrics: the best performance model was LightGBM. Logistic regression with three variables (LR3) selected via bootstrapping showed performances similar to the best-performing models. The AUC of test data is slightly above 0.65 for the best models (highest value: 0.662 with LightGBM). Conclusions. No model performed the best for all metrics: more complex ML models had better performances; however, models with just three features showed performances comparable to the best models using many (n = 13–19) features. [ABSTRACT FROM AUTHOR]

Purpose: The purpose of this study is to study the evolution of quality of life (QoL) in the first 5 years following Intensity-modulated radiation therapy (IMRT) for prostate cancer (PCa) and to determine possible associations with clinical/treatment data. Material and methods: Patients were enrolled in a prospective multicentre observational trial in 2010-2014 and treated with conventional (74–80 Gy, 1.8–2 Gy/fr) or moderately hypofractionated IMRT (65–75.2 Gy, 2.2–2.7 Gy/fr). QoL was evaluated by means of EORTC QLQ-C30 at baseline, at radiation therapy (RT) end, and every 6 months up to 5 years after IMRT end. Fourteen QoL dimensions were investigated separately. The longitudinal evaluation of QoL was analysed by means of Analysis of variances (ANOVA) for multiple measures. Results: A total of 391 patients with complete sets of questionnaires across 5 years were available. The longitudinal analysis showed a trend toward the significant worsening of QoL at RT end for global health, physical and role functioning, fatigue, appetite loss, diarrhoea, and pain. QoL worsening was recovered within 6 months from RT end, with the only exception being physical functioning. Based on ANOVA, the most impaired time point was RT end. QoL dimension analysis at this time indicated that acute Grade ≥ 2 gastrointestinal (GI) toxicity significantly impacted global health, physical and role functioning, fatigue, appetite loss, diarrhoea, and pain. Acute Grade ≥ 2 genitourinary (GU) toxicity resulted in lower role functioning and higher pain. Prophylactic lymph-nodal irradiation (WPRT) resulted in significantly lower QoL for global health, fatigue, appetite loss, and diarrhoea; lower pain with the use of neoadjuvant/concomitant hormonal therapy; and lower fatigue with the use of an anti-androgen. Conclusions: In this prospective, longitudinal, observational study, high radiation IMRT doses delivered for PCa led to a temporary worsening of QoL, which tended to be completely resolved at six months. Such transient worsening was mostly associated with acute GI/GU toxicity, WPRT, and higher prescription doses. [ABSTRACT FROM AUTHOR]

Stability analysis remains a fundamental step in developing a successful imaging biomarker to personalize oncological strategies. This study proposes an in silico contour generation method for simulating segmentation variations to identify stable radiomic features. Ground-truth annotation provided for the whole prostate gland on the multi-parametric MRI sequences (T2w, ADC, and SUB-DCE) were perturbed to mimic segmentation differences observed among human annotators. In total, we generated 15 synthetic contours for a given image-segmentation pair. One thousand two hundred twenty-four unfiltered/filtered radiomic features were extracted applying Pyradiomics, followed by stability assessment using ICC(1,1). Stable features identified in the internal population were then compared with an external population to discover and report robust features. Finally, we also investigated the impact of a wide range of filtering strategies on the stability of features. The percentage of unfiltered (filtered) features that remained robust subjected to segmentation variations were T2w—36% (81%), ADC—36% (94%), and SUB—43% (93%). Our findings suggest that segmentation variations can significantly impact radiomic feature stability but can be mitigated by including pre-filtering strategies as part of the feature extraction pipeline. [ABSTRACT FROM AUTHOR]

Simple Summary: Despite the implementation of specific dose constraints on healthy tissues to maintain the theoretical risk of late toxicity below 5% five years after radiotherapy (RT), many patients experience "unusual" toxicity during their oncological follow-up. This narrative review describes the methods of individual radiation sensitivity (iRS) diagnosis, their impact on the RT workflow as well as initiatives to support clinical decision-making initiatives. (1) Background: radiotherapy is a cornerstone of cancer treatment. When delivering a tumoricidal dose, the risk of severe late toxicities is usually kept below 5% using dose-volume constraints. However, individual radiation sensitivity (iRS) is responsible (with other technical factors) for unexpected toxicities after exposure to a dose that induces no toxicity in the general population. Diagnosing iRS before radiotherapy could avoid unnecessary toxicities in patients with a grossly normal phenotype. Thus, we reviewed iRS diagnostic data and their impact on decision-making processes and the RT workflow; (2) Methods: following a description of radiation toxicities, we conducted a critical review of the current state of the knowledge on individual determinants of cellular/tissue radiation; (3) Results: tremendous advances in technology now allow minimally-invasive genomic, epigenetic and functional testing and a better understanding of iRS. Ongoing large translational studies implement various tests and enriched NTCP models designed to improve the prediction of toxicities. iRS testing could better support informed radiotherapy decisions for individuals with a normal phenotype who experience unusual toxicities. Ethics of medical decisions with an accurate prediction of personalized radiotherapy's risk/benefits and its health economics impact are at stake; (4) Conclusions: iRS testing represents a critical unmet need to design personalized radiotherapy protocols relying on extended NTCP models integrating iRS. [ABSTRACT FROM AUTHOR]

Prostate cancer (PCa) ranges from indolent to aggressive tumors that may rapidly progress and metastasize. The switch to aggressive PCa is fostered by reactive stroma infiltrating tumor foci. Therefore, reactive stroma-based biomarkers may potentially improve the early detection of aggressive PCa, ameliorating disease classification. Gene expression profiles of PCa reactive fibroblasts highlighted the up-regulation of genes related to stroma deposition, including periostin and sparc. Here, the potential of periostin as a stromal biomarker has been investigated on PCa prostatectomies by immunohistochemistry. Moreover, circulating levels of periostin and sparc have been assessed in a low-risk PCa patient cohort enrolled in active surveillance (AS) by ELISA. We found that periostin is mainly expressed in the peritumoral stroma of prostatectomies, and its stromal expression correlates with PCa grade and aggressive disease features, such as the cribriform growth. Moreover, stromal periostin staining is associated with a shorter biochemical recurrence-free survival of PCa patients. Interestingly, the integration of periostin and sparc circulating levels into a model based on standard clinico-pathological variables improves its performance in predicting disease reclassification of AS patients. In this study, we provide the first evidence that circulating molecular biomarkers of PCa stroma may refine risk assessment and predict the reclassification of AS patients. [ABSTRACT FROM AUTHOR]

Simple Summary: Radiotherapy is the cornerstone of treatment of nasopharyngeal cancer, in different settings with or without chemotherapy. This role has been recently strengthened by the introduction of proton therapy, as a radiation treatment option for head and neck cancer, obtaining improved plans with a reduced dose to organs-at-risk. Definition of strategies to identify patients who would benefit the most from proton therapy in terms of reduced toxicity is highly desirable, due to limited availability and higher costs of this treatment option. Two parallel working pipelines were depicted in this study for nasopharyngeal cancer patients. The introduction of a synthetic index describing the overall expected reduction in toxicities in the head and neck region with proton therapy was supported by the application of the well-established model-based selection methodology, relative to the same patient cohort. Based on this analysis, the fraction of nasopharyngeal cancer patients expected to receive a benefit with proton therapy was in line with the Dutch experience for the head and neck cancer population. (1) Background: we proposed an integrated strategy to support clinical allocation of nasopharyngeal patients between proton and photon radiotherapy. (2) Methods: intensity-modulated proton therapy (IMPT) plans were optimized for 50 consecutive nasopharyngeal carcinoma (NPC) patients treated with volumetric modulated arc therapy (VMAT), and differences in dose and normal tissue complication probability (ΔNTCPx-p) for 16 models were calculated. Patient eligibility for IMPT was assessed using a model-based selection (MBS) strategy following the results for 7/16 models describing the most clinically relevant endpoints, applying a model-specific ΔNTCPx-p threshold (15% to 5% depending on the severity of the complication) and a composite threshold (35%). In addition, a comprehensive toxicity score (CTS) was defined as the weighted sum of all 16 ΔNTCPx-p, where weights follow a clinical rationale. (3) Results: Dose deviations were in favor of IMPT (ΔDmean ≥ 14% for cord, esophagus, brainstem, and glottic larynx). The risk of toxicity significantly decreased for xerostomia (−12.5%), brain necrosis (−2.3%), mucositis (−3.2%), tinnitus (−8.6%), hypothyroidism (−9.3%), and trismus (−5.4%). There were 40% of the patients that resulted as eligible for IMPT, with a greater advantage for T3–T4 staging. Significantly different CTS were observed in patients qualifying for IMPT. (4) Conclusions: The MBS strategy successfully drives the clinical identification of NPC patients, who are most likely to benefit from IMPT. CTS summarizes well the expected global gain. [ABSTRACT FROM AUTHOR]

Simple Summary: We built a predictive model for acute salivary dysfunction for nasopharyngeal cancer patients receiving combined treatment. The final aim was to provide a nomogram (with dosimetric and clinical risk factors) to help physicians in the streamline prevention and management of this acute side effect. No research has focused on predicting acute xerostomia so far. We do not know if models predicting late xerostomia can also be applied to acute xerostomia, since different pathogenesis is suggested for acute and late events. The model was tested in two independent external cohorts. Validation results highlighted that the dosimetric part of the predictive model was highly generalisable, with the clinical risk part still being a weak component. The good validation of the model's discriminative power and of the effect of the size of dosimetric factors created confidence for considering these factors while optimising radiotherapy. Background: Radiation-induced xerostomia is one of the most prevalent adverse effects of head and neck cancer treatment, and it could seriously affect patients' qualities of life. It results primarily from damage to the salivary glands, but its onset and severity may also be influenced by other patient-, tumour-, and treatment-related factors. We aimed to build and validate a predictive model for acute salivary dysfunction (aSD) for locally advanced nasopharyngeal carcinoma (NPC) patients by combining clinical and dosimetric factors. Methods: A cohort of consecutive NPC patients treated curatively with IMRT and chemotherapy at 70 Gy (2–2.12 Gy/fraction) were utilised. Parotid glands (cPG, considered as a single organ) and the oral cavity (OC) were selected as organs-at-risk. The aSD was assessed at baseline and weekly during RT, grade ≥ 2 aSD chosen as the endpoint. Dose-volume histograms were reduced to the Equivalent Uniform Dose (EUD). Dosimetric and clinical/treatment features selected via LASSO were inserted into a multivariable logistic model. Model validation was performed on two cohorts of patients with prospective aSD, and scored using the same schedule/scale: a cohort (NPC_V) of NPC patients (as in model training), and a cohort of mixed non-NPC head and neck cancer patients (HNC_V). Results: The model training cohort included 132 patients. Grade ≥ 2 aSD was reported in 90 patients (68.2%). Analyses resulted in a 4-variables model, including doses of up to 98% of cPG (cPG_D98%, OR = 1.04), EUD to OC with n = 0.05 (OR = 1.11), age (OR = 1.08, 5-year interval) and smoking history (OR = 1.37, yes vs. no). Calibration was good. The NPC_V cohort included 38 patients, with aSD scored in 34 patients (89.5%); the HNC_V cohort included 93 patients, 77 with aSD (92.8%). As a general observation, the incidence of aSD was significantly different in the training and validation populations (p = 0.01), thus impairing calibration-in-the-large. At the same time, the effect size for the two dosimetric factors was confirmed. Discrimination was also satisfactory in both cohorts: AUC was 0.73, and 0.68 in NPC_V and HNC_V cohorts, respectively. Conclusion: cPG D98% and the high doses received by small OC volumes were found to have the most impact on grade ≥ 2 acute xerostomia, with age and smoking history acting as a dose-modifying factor. Findings on the development population were confirmed in two prospectively collected validation populations. [ABSTRACT FROM AUTHOR]

Simple Summary: Dosiomics is born directly as an extension of radiomics: it entails extracting features from the patients' three-dimensional (3D) radiotherapy dose distribution rather than from conventional medical images to obtain specific spatial and statistical information. Dosiomic studies, in a multicentre setting, require assessing the features' stability to dose calculation settings and the features' capability in distinguishing different dose distributions. This study provides the first multicentre evaluation of the dosiomic features in terms of reproducibility, stability and sensitivity across various dose distributions obtained from multiple technologies and techniques and considering different dose calculation algorithms of TPS and two different resolutions of the dose grid. Harmonisation strategies to account for a possible variation in the dose distribution due to these confounding factors should be adopted when investigating a correlation between dosiomic features and clinical outcomes in multicentre studies. Dosiomics is a texture analysis method to produce dose features that encode the spatial 3D distribution of radiotherapy dose. Dosiomic studies, in a multicentre setting, require assessing the features' stability to dose calculation settings and the features' capability in distinguishing different dose distributions. Dose distributions were generated by eight Italian centres on a shared image dataset acquired on a dedicated phantom. Treatment planning protocols, in terms of planning target volume coverage and dose–volume constraints to the organs at risk, were shared among the centres to produce comparable dose distributions for measuring reproducibility/stability and sensitivity of dosiomic features. In addition, coefficient of variation (CV) was employed to evaluate the dosiomic features' variation. We extracted 38,160 features from 30 different dose distributions from six regions of interest, grouped by four features' families. A selected group of features (CV < 3 for the reproducibility/stability studies, CV > 1 for the sensitivity studies) were identified to support future multicentre studies, assuring both stable features when dose distributions variation is minimal and sensitive features when dose distribution variations need to be clearly identified. Dosiomic is a promising tool that could support multicentre studies, especially for predictive models, and encode the spatial and statistical characteristics of the 3D dose distribution. [ABSTRACT FROM AUTHOR]

Simple Summary: The levels of urinary incontinence (UI) at adjuvant/salvage radiotherapy (ART/SRT) start strongly influence long-term UI recovery, possibly inducing clinicians to postpone radiotherapy "as much as possible" in order to maximize UI recovery, but possibly reducing radiotherapy efficacy. Our study analyzed UI recovery from prostatectomy to ART/SRT by means of the International Consultation on Incontinence Questionnaire-Short Form filled-in by patients at ART/SRT start. Three endpoints were investigated: frequency and amount of urine loss and the "subjective" patient-perceived detrimental impact on quality of life, as well as the possible influence of clinical and personality variables. The time elapsed from prostatectomy to radiotherapy start was the strongest predictor for each UI endpoint, all improving between four and eight months after prostatectomy, but without any additional long-term recovery. Background: Baseline urinary incontinence (UI) strongly modulates UI recovery after adjuvant/salvage radiotherapy (ART/SRT), inducing clinicians to postpone it "as much as possible", maximizing UI recovery but possibly reducing efficacy. This series aims to analyze the trend of UI recovery and its predictors at radiotherapy start. Methods: A population of 408 patients treated with ART/SRT enrolled in a cohort study (ClinicalTrials.gov #NCT02803086) aimed at developing predictive models of radiation-induced toxicities. Self-reported UI and personality traits, evaluated by means of the International Consultation on Incontinence Questionnaire-Urinary Incontinence Short Form (ICIQ-SF) and Eysenck Personality Questionnaire - Revised (EPQ-R) questionnaires, were assessed at ART/SRT start. Several endpoints based on baseline ICIQ-SF were investigated: frequency and amount of urine loss (ICIQ3 and ICIQ4, respectively), "objective" UI (ICIQ3 + 4), "subjective" UI (ICIQ5), and "TOTAL" UI (ICIQ3 +4 + 5). The relationship between each endpoint and time from prostatectomy to radiotherapy (TTRT) was investigated. The association between clinical and personality variables and each endpoint was tested by uni- and multivariable logistic regression. Results: TTRT was the strongest predictor for all endpoints (p-values ≤ 0.001); all scores improved between 4 and 8 months after prostatectomy, without any additional long-term recovery. Neuroticism independently predicted subjective UI, TOTAL UI, and daily frequency. Conclusions: Early UI recovery mostly depends on TTRT with no further improvement after 8 months from prostatectomy. Higher levels of neuroticism may overestimate UI. [ABSTRACT FROM AUTHOR]

Simple Summary: Active surveillance (AS) has evolved as an alternative to radical treatment for potentially indolent prostate cancer. However, current selection criteria for entering AS are suboptimal, and a significant percentage of patients discontinue AS because of disease reclassification. Hence, there is an unmet need for novel biomarkers for the accurate identification of high-risk PCa and the unequivocal classification of indolent disease. Circulating biomarkers, including microRNAs identified through liquid biopsies, represent a valuable approach to improve on currently available clinicopathological risk-stratification tools. In an attempt to identify specific microRNA signatures as potential circulating biomarkers, the authors performed an unprecedented analysis of the global microRNA profile in plasma samples from AS patients and identified and validated a three-microRNA signature able to predict patient reclassification. The addition of the three-microRNA signature was able to improve the performance of currently available clinicopathological variables, thus showing potential for the refinement of AS patients' selection. Active surveillance (AS) has evolved as a strategy alternative to radical treatments for very low risk and low-risk prostate cancer (PCa). However, current criteria for selecting AS patients are still suboptimal. Here, we performed an unprecedented analysis of the circulating miRNome to investigate whether specific miRNAs associated with disease reclassification can provide risk refinement to standard clinicopathological features for improving patient selection. The global miRNA expression profiles were assessed in plasma samples prospectively collected at baseline from 386 patients on AS included in three independent mono-institutional cohorts (training, testing and validation sets). A three-miRNA signature (miR-511-5p, miR-598-3p and miR-199a-5p) was found to predict reclassification in all patient cohorts (training set: AUC 0.74, 95% CI 0.60–0.87, testing set: AUC 0.65, 95% CI 0.51–0.80, validation set: AUC 0.68, 95% CI 0.56–0.80). Importantly, the addition of the three-miRNA signature improved the performance of the clinical model including clinicopathological variables only (AUC 0.70, 95% CI 0.61–0.78 vs. 0.76, 95% CI 0.68–0.84). Overall, we trained, tested and validated a three-miRNA signature which, combined with selected clinicopathological variables, may represent a promising biomarker to improve on currently available clinicopathological risk stratification tools for a better selection of truly indolent PCa patients suitable for AS. [ABSTRACT FROM AUTHOR]

Simple Summary: Ionizing radiation is used as a treatment for cancer, but it also affects the endothelial cells that make up the microvasculature. In-vitro models can be used to study the detrimental effect of irradiation on those cells. This systematic review analyzed the literature models, highlighting the critical components of the production, irradiation, and analysis of radiobiological in-vitro models for microvascular endothelial cell damage. Based on those data, we suggest future directions, including advanced in-vitro models to recapitulate microenvironment features. We pinpoint essential information to be included for the good characterization of the experiments, especially in terms of the dose delivered by ionizing radiation. Ionizing radiation (IR) is used in radiotherapy as a treatment to destroy cancer. Such treatment also affects other tissues, resulting in the so-called normal tissue complications. Endothelial cells (ECs) composing the microvasculature have essential roles in the microenvironment's homeostasis (ME). Thus, detrimental effects induced by irradiation on ECs can influence both the tumor and healthy tissue. In-vitro models can be advantageous to study these phenomena. In this systematic review, we analyzed in-vitro models of ECs subjected to IR. We highlighted the critical issues involved in the production, irradiation, and analysis of such radiobiological in-vitro models to study microvascular endothelial cells damage. For each step, we analyzed common methodologies and critical points required to obtain a reliable model. We identified the generation of a 3D environment for model production and the inclusion of heterogeneous cell populations for a reliable ME recapitulation. Additionally, we highlighted how essential information on the irradiation scheme, crucial to correlate better observed in vitro effects to the clinical scenario, are often neglected in the analyzed studies, limiting the translation of achieved results. [ABSTRACT FROM AUTHOR]

Simple Summary: The prognostic performance of traditional methodologies in advanced nasopharyngeal carcinoma does not allow to successfully stratify patients. Previous studies showed that MRI-radiomics has been used to give additional information to improve the prognosis for this type of pathology in patients from endemic areas (Asia). The purpose of this study was to use MRI-radiomics to develop prognostic models for overall survival in patients from non-endemic areas (Europe or United States). In particular, T1-weighted and T2-weighted MRI were used for the purpose. Radiomic features from those images allowed to successfully train a prognostic signature that improved the prognostic performance of models based on clinical variables alone for different clinical endpoints (overall survival, disease-free survival and loco-regional recurrence-free survival). These results suggest how MRI-radiomics is a useful additional tool for prognosis in nasopharyngeal cancer. Advanced stage nasopharyngeal cancer (NPC) shows highly variable treatment outcomes, suggesting the need for independent prognostic factors. This study aims at developing a magnetic resonance imaging (MRI)-based radiomic signature as a prognostic marker for different clinical endpoints in NPC patients from non-endemic areas. A total 136 patients with advanced NPC and available MRI imaging (T1-weighted and T2-weighted) were selected. For each patient, 2144 radiomic features were extracted from the main tumor and largest lymph node. A multivariate Cox regression model was trained on a subset of features to obtain a radiomic signature for overall survival (OS), which was also applied for the prognosis of other clinical endpoints. Validation was performed using 10-fold cross-validation. The added prognostic value of the radiomic features to clinical features and volume was also evaluated. The radiomics-based signature had good prognostic power for OS and loco-regional recurrence-free survival (LRFS), with C-index of 0.68 and 0.72, respectively. In all the cases, the addition of radiomics to clinical features improved the prognostic performance. Radiomic features can provide independent prognostic information in NPC patients from non-endemic areas. [ABSTRACT FROM AUTHOR]