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]

Abstract Background To analyse limits and capabilities in dose calculation of collapsed-cone-convolution (CCC) algorithm implemented in helical tomotherapy (HT) treatment planning system for thorax lesions. Methods The agreement between measured and calculated dose was verified both in homogeneous (Cheese Phantom) and in a custom-made inhomogeneous phantom. The inhomogeneous phantom was employed to mimic a patient's thorax region with lung density encountered in extreme cases and acrylic inserts of various dimensions and positions inside the lung cavity. For both phantoms, different lung treatment plans (single or multiple metastases and targets in the mediastinum) using HT technique were simulated and verified. Point and planar dose measurements, both with radiographic extended-dose-range (EDR2) and radiochromic external-beam-therapy (EBT2) films, were performed. Absolute point dose measurements, dose profile comparisons and quantitative analysis of gamma function distributions were analyzed. Results An excellent agreement between measured and calculated dose distributions was found in homogeneous media, both for point and planar dose measurements. Absolute dose deviations Conclusions Very acceptable accuracy was found for complex lung treatment plans calculated with CCC algorithm implemented in the tomotherapy TPS even in the heterogeneous phantom with very low lung-density.

Objective: To evaluate toxicity and outcomes after salvage robotic stereotactic body radiotherapy-SBRT- (CyberKnife®,Accuray, Sunnyvale, Ca) re-irradiation of patients with intraprostatic/prostatic bed recurrences of pelvic malignancies. Methods: From 11/2018 to 10/2021, 20 patients with intraprostatic/prostatic bed recurrence after radiotherapy, diagnosed on MRI/ PET choline/PSMA and/or biopsyproven, underwent a salvage re-irradiation with SBRT: 9 patients with prostate cancer failure after previous radical radiotherapy, and 11 patients after adjuvant/salvage radiotherapy. Median prior RT dose was 70.7 (38.25- 78) Gy and the median interval to SBRT salvage therapy was 79.3 (11-208) months. Median PSA before robotic SBRT was 2.64 (1.14-26.8) ng/ml. Fiducial markers were implanted into the target in 19 of 20 patients. Median SBRT total dose was 35 (30-35) Gy in 5 fractions (EQD2=85 Gy, for α/β 1.5). Median prescription isodose was 70% (59- 81%). In 12 patients, a “urethral sparing” was used. In 10 cases a precautionary therapy with steroids and alphalytics was prescribed during the salvage treatment. Twelve patients received neoadjuvant or concomitant/adjuvant androgen suppressive therapy during their SBRT course. Toxicity was scored in accordance with CTCAE v 5.0. Results: Median follow-up was 17.4 months (2.35- 38.15) months. Acute genitourinary (GU) toxicity was observed in 40% of patients and was limited to grade (G)1 in 35% of patients and G2 (minimal strangury, urgency and occasional urinary incontinence) in 5%. Ten% of patients reported late GU G3 toxicity (urinary retention requiring catheterization and transurethral resection), and 5% G2 urinary toxicity. No acute and late gastrointestinal toxicity was observed. At the last follow-up 2 patients died due to a non cancer-releated cause. Twenty-four-months Kaplan Meier estimates of biochemical relapse-free survival (bRFS) was 78.2% (see Figure 1), local-relapse-free survival (LRFS) 92.3% and distant-metastases-free survival 85.1%. Conclusions: Prostate SBRT re-irradiation with CK is a feasible treatment option, with good short-term outcomes. Longer follow-up is necessary to assess the long-term benefits and to determine late toxicity. [ABSTRACT FROM AUTHOR]

Aims: This study reports the preliminary results of a monoinstitutional experience with stereotactic body radiotherapy (SBRT) performed on oligometastatic gynecological cancer patients. Methods: From 04/2009 to 07/2021, 85 lesions (46 patients) were treated: 5 lesions (4 patients) with Image Guided-helical Intensity Modulated Radiotherapy (IGIMRT) at a median dose of 54 (35-63) Gy in 6 (5-10) median fractions prescribed at 95% of the Planning Target Volume (PTV), and 80 lesions (42 patients) with robotic SBRT at a median dose of 40 (18-60) Gy in 5 (1- 8) fractions, prescribed at a median isodose of 80% (68- 84%). Ten.6% of lesions were in the field of a previous adjuvant/salvage IG-IMRT performed at a median dose of 50.4 Gy. Primary origin was: uterine in 39.1%, ovarian/ fallopian tubes in 34.8%, and cervical/vulvar/vaginal in 26.1% of patients. Metastatic site was: brain in 6.5%, liver in 4.3%, lymph-nodes in 60.9%, lung in 23.9%, and bone in 4.3% of patients. Number of concomitant lesions was 1 in 67.4%, 2 in 19.6%, 3 in 2.12%, 4 in 4.3%, and 5 in 6.5% of patients. De nuovo oligometastatic were 32.6%, induced 23.9%, and relapsed 43.5% of patients. Results: Median age at the treatment was 67.4 years (Interquartile Range, IQR: 58.6-75.7). Median follow-up from initial diagnosis was 75.6 months (IQR:46.4–164.8). Median Biologically Effective Dose(BED) prescribed: 79.2 Gy (IQR:59.5-102.6). Twelve and 24 month local relapsefree survival(LRFS) was 88.3 and 84.3%, respectively, while 12 and 24 month distant relapse-free survival(DRFS) was 44.6 and 27.6%. Twenty-four and 36 months overall survival (OS) was 100% and 85% respectively. Acute toxicity registered was grade (G) 1 13% and G2 6.5%. Late toxicity was 2.2% G1, 2.2% G2 and 2.2% (1 patient) G3 (vertebral fracture). There were no differences in DRFS (see Fig.1) and OS between the three groups. At the multivariate analysis factors predicting distant relapse were: age at diagnosis HR=1.07 (95%CI:1.03-1.13, p=0.0021), number of lesions≥3 vs 1, HR= 4.16 (95%CI:1.34-12.9, p=0.0135), and induced vs de nuovo oligometastatic lesions HR=8.55 (95%CI:1.96-37.3, p=0.0043), but not relapsed vs de nuovo(p=0.067). Conclusion: SBRT is effective for the local control of oligometastatic disease from gynecological tumors, with low toxicity. Distant relapse remains the primary cause of failure after SBRT in oligometastatic gynecological patients, and a number of lesions≥3, age at diagnosis and induced oligometastatic lesions are predictive for distant relapse. [ABSTRACT FROM AUTHOR]

Objective: We analyzed initial outcomes and toxicities of the 2 schedules of stereotactic body radiotherapy (SBRT) in use in our department for prostate cancer (PCa) patients. Materials and methods: Between 10/2017-05/2021, 102 PCa patients were treated using robotic SBRT with two different schedules. Group I included 73 patients treated at 36.25 Gy/ 5 fractions at the 79% median isodose, group II 29 patients treated at 38 Gy/ 4 fractions at the 63% median isodose (urethral sparing HDR-like technique). Median age was 74.5 vs 75.8 years, median initial PSA 7.20 vs 7.65 ng/ml. Gleason score was 3+3 in 11% vs 17.2%, 3+4 in 58.9% vs 51.7%, 4+3 in 17.8% vs 20.7% and not accorded (due to previous androgen deprivation) in 1.4% vs 3.5% of patients. Androgen deprivation therapy (ADT) was prescribed in 45% vs 31% of patients, with a median duration of 7 vs 6 months. Fiducial markers were implanted into the prostate in all patients. In 52% of patients steroid therapy and/or alpha-lytics were prescribed to prevent side effects. Toxicity was scored in accordance with CTCAE v 5.0. Biochemical failure was assessed using the nadir + 2 definition. Results: Median follow-up was 23.4 (0-47.2) months in group I vs 12 (2.78-36.07) months in group II. For acute and late toxicities see Table 1. No acute grade(G) 3 toxicities were registered. One (1.4%) late G3 genitourinary toxicity (transurethral incision) occurred in a patient treated with 36.25 Gy. Median post-SBRT PSA level was 0.342 (0.001-231.00) ng/ml in group I and 0.603 (0.05-10.210) ng/ml in group II. At the last follow up 3 pts had died due to a non-cancer related cause, 4 pts had a biochemical failure with PSMA/CholinePET positive progression: 3 treated at 36.25 Gy and 1 treated at 38 Gy. Two-year biochemical relapse-free survival(bRFS) was 94.8% for the patients of group I and 88.9% for the patients of group II, respectively (p= 0.6). Overall survival (OS) was 85.7%vs 98.4% at 2 years(p=0.35). Disease-free survival(DFS) was 94.8% and 88.9%, respectively (p=0.6). No correlation was found between target volume and toxicity. Conclusion: The urethral sparing technique allowed dose escalation (from 2 Gy equivalent dose-EQD2-91 Gy to EQD2 ≥ 120 Gy with α/β 1.5) without increasing G3 toxicity and with non-inferior bRFS despite less ADT prescription. Longer follow-up is needed to confirm these results. [ABSTRACT FROM AUTHOR]