Fit of biokinetic data in molecular radiotherapy: a machine learning approach.

BACKGROUND: In literature are reported different analytical methods (AM) to choose the proper fit model and to fit data of the time-activity curve (TAC). On the other hand, Machine Learning algorithms (ML) are increasingly used for both classification and regression tasks. The aim of this work was to investigate the possibility of employing ML both to classify the most appropriate fit model and to predict the area under the curve (τ). METHODS: Two different ML systems have been developed for classifying the fit model and to predict the biokinetic parameters. The two systems were trained and tested with synthetic TACs simulating a whole-body Fraction Injected Activity for patients affected by metastatic Differentiated Thyroid Carcinoma, administered with [131I]I-NaI. Test performances, defined as classification accuracy (CA) and percentage difference between the actual and the estimated area under the curve (Δτ), were compared with those obtained using AM varying the number of points (N) of the TACs. A comparison between AM and ML were performed using data of 20 real patients. RESULTS: As N varies, CA remains constant for ML (about 98%), while it improves for F-test (from 62 to 92%) and AICc (from 50 to 92%), as N increases. With AM, [Formula: see text] can reach down to - 67%, while using ML [Formula: see text] ranges within ± 25%. Using real TACs, there is a good agreement between τ obtained with ML system and AM. CONCLUSIONS: The employing of ML systems may be feasible, having both a better classification and a better estimation of biokinetic parameters.

Dose-Related Analysis in Percutaneous Central Venous Catheters Insertion: Experience of a Pediatric Interventional Radiology Center.

BACKGROUND: There are many techniques for long-term central venous catheter (CVC) placement, but none of them are specific for pediatric patients or focused on the delivered dose of ionizing radiation. MATERIALS AND METHODS: This retrospective study examined a sample of pediatric patients who received percutaneous long-term CVC positioning in a tertiary care pediatric hospital. Effective dose, dose-area product (DAP) and length of time of exposition during the procedure were determined, using an appropriate technical procedure, exam and program set of the angiograph, and compared with an unpaired t-test analysis. RESULTS: The study included 1410 enrolled patients, with a median age of 10 years (range 0.2-18 years), between 2016 and 2019. In 2016 (318 pts), the mean effective dose was 0.13 mSv and the mean DAP dose was 18.95 µGy/m2 In 2017 (353 pts), the mean effective dose was 0.11 mSv and the mean DAP dose was 17.26 µGy/m2. In 2018 (351 pts), the mean effective dose was 0.05 mSv and the mean DAP dose was 7.23 µGy/m2. In 2019 (388 pts), the mean effective dose was 0.02 mSv and the mean DAP dose was 3.10 µGy/m2. CONCLUSIONS: Medical and technical expertise led to a remarkable reduction in the radiation dose. Therefore, the authors' hypothesis is that US- and fluoroscopy-guided percutaneous long-term CVC insertion technique is safer, more cost-effective and lower in terms of radiation exposure if correctly applied, compared to surgical or percutaneous by direct puncture techniques.

Personalized dosimetry for a deeper understanding of metastatic response to high activity 131I-mIBG therapy in high risk relapsed refractory neuroblastoma.

BACKGROUND: Dosimetry in molecular radiotherapy for personalized treatment is assuming a central role in clinical management of aggressive/relapsed tumors. Relapsed/refractory metastatic high-risk neuroblastoma (rrmHR-NBL) has a poor prognosis and high-activity 131I-mIBG therapy could represent a promising strategy. The primary aim of this case series study was to report the absorbed doses to whole-body (DWB ), red marrow (DRM ) and lesions (DLesion ). A secondary aim was to correlate DLesion values to clinical outcome. METHODS: Fourteen patients affected by rrmHR-NBL were treated with high-activity 131I-mIBG therapy (two administrations separated by 15 days). The first administration was weight-based whereas the second one was dosimetry-based (achieving DWB equals to 4 Gy). In all patients DWB and DRM was assessed; 9/14 patients were selected for DLesion evaluation using planar dosimetric approach (13 lesions evaluated). Treatment response was classified as progressive and stable disease (PD and SD), partial and complete response (PR and CR) according to the International Neuroblastoma Response Criteria. Patients were divided into two groups: Responder (CR, PR, SD) and Non-Responder (PD), correlating treatment response to DLesion value. RESULTS: The cumulative DWB , DRM and DLesion ranged from (1.5; 4.5), (1.0; 2.6) and (44.2; 585.8) Gy. A linear correlation between DWB and DRM and a power law correlation between the absorbed dose to WB normalized for administered activity and the mass of the patient were observed. After treatment 3, 2, 4 and 5 patients showed CR, PR, SD and PD respectively, showing a correlation between DLesion and the two response group. CONCLUSIONS: Our experience demonstrated feasibility of high activity therapy of 131I-mIBG in rrmHR-NBL children as two administration intensive strategy. Dosimetric approach allowed a tailored high dose treatment maximizing the benefits of radionuclide therapy for pediatric patients while maintaining a safety profile. The assesment of DLesion contributed to have a deeper understaning of metabolic treatment effects.

Scintigraphic load of bone disease evaluated by DASciS software as a survival predictor in metastatic castration-resistant prostate cancer patients candidates to 223RaCl treatment.

Background Aim of our study was to assess the load of bone disease at starting and during Ra-223 treatment as an overall survival (OS) predictor in metastatic castration-resistant prostate cancer (mCRPC) patients. Bone scan index (BSI) is defined as the percentage of total amount of bone metastasis on whole-body scintigraphic images. We present a specific software (DASciS) developed by an engineering team of "Sapienza" University of Rome for BSI calculation. Patients and methods 127 mCRPC patients bone scan images were processed with DASciS software, and BSI was tested as OS predictor. Results 546 bone scans were analyzed revealing that the extension of disease is a predictor of OS (0-3% = 28 months of median survival (MoMS]; 3%-5% = 11 MoMS, > 5% = 5 MoMS). BSI has been analyzed as a single parameter for OS, determining an 88% AUC. Moreover, the composition between the BSI and the 3-PS (3-variable prognostic score) determines a remarkable improvement of the AUC (91%), defining these two parameters as the best OS predictors. Conclusions This study suggests that OS is inversely correlated with the load of bone disease in mCRPC Ra-223-treated subjects. DASciS software appears a promising tool in identifying mCRPC patients that more likely take advantage from Ra-223 treatment. BSI is proposed as a predictive variable for OS and included to a multidimensional clinical evaluation permits to approach the patients' enrollment in a rational way, allowing to enhance the treatment effectiveness together with cost optimization.