The role of the deep convolutional neural network as an aid to interpreting brain [18F]DOPA PET/CT in the diagnosis of Parkinson's disease.

OBJECTIVES: To test the performance of a 3D convolutional neural network (CNN) in analysing brain [18F]DOPA PET/CT in order to identify patients with nigro-striatal neurodegeneration. We evaluated the robustness of the 3D CNN by testing it against a manual regional analysis of the striata by using a striatal-to-occipital ratio (SOR). METHODS: We analyzed patients who had undergone [18F]DOPA PET/CT from 2016 to 2018. Two examiners interpreted PET/CT images as positive or negative. Only patients with at least 2 years of follow-up and an ascertained neurological diagnosis were included. A 3D CNN was developed to evaluate [18F]DOPA PET/CT and refine the diagnosis of movement disorder. This system required training and testing, which were carried out on 2/3 and 1/3 of patients, respectively. A regional analysis was also conducted by drawing region of interest on T1-weighted 3D MRI scans, on which the [18F]DOPA PET images were first co-registered. RESULTS: Ninety-eight patients were enrolled: 43 presented nigro-striatal degeneration and 55 negative cases used as controls. After training on 69 patients, the diagnostic performance of the 3D CNN was then calculated in 29 patients. Sensitivity, specificity, negative predictive value, positive predictive value and accuracy were 100%, 89%, 100%, 85% and 93%, respectively. When we compared the 3D CNN results with the SOR analysis, we found that the two patients falsely classified as positive by the 3D CNN procedure showed SOR values ≤ 5th percentile of the negative cases' distribution. CONCLUSIONS: 3D CNNs are able to interpret [18F]DOPA PET/CT properly, revealing patients affected by Parkinson's disease. KEY POINTS: • [18F]DOPA PET/CT is a sensitive diagnostic tool to identify patients with nigro-striatal neurodegeneration. • A semiquantitative evaluation of the images allows a more confident interpretation of the PET findings. • 3D convolutional neural network allows an accurate interpretation of 18F-DOPA PET/CT images, revealing patients affected by Parkinson's disease.

The influence of small field output factors simulated uncertainties on the calculated dose in VMAT plans for brain metastases: a multicentre study.

OBJECTIVES: This multicentric study was carried out to investigate the impact of small field output factors (OFs) inaccuracies on the calculated dose in volumetric arctherapy (VMAT) radiosurgery brain plans. METHODS: Nine centres, realised the same five VMAT plans with common planning rules and their specific clinical equipment Linac/treatment planning system commissioned with their OFs measured values (OFbaseline). In order to simulate OFs errors, two new OFs sets were generated for each centre by changing only the OFs values of the smallest field sizes (from 3.2 × 3.2 cm2 to 1 × 1 cm2) with well-defined amounts (positive and negative). Consequently, two virtual machines for each centre were recommissioned using the new OFs and the percentage dose differences ΔD (%) between the baseline plans and the same plans recalculated using the incremented (OFup) and decremented (OFdown) values were evaluated. The ΔD (%) were analysed in terms of planning target volume (PTV) coverage and organs at risk (OARs) sparing at selected dose/volume points. RESULTS: The plans recalculated with OFdown sets resulted in higher variation of doses than baseline within 1.6 and 3.4% to PTVs and OARs respectively; while the plans with OFup sets resulted in lower variation within 1.3% to both PTVs and OARs. Our analysis highlights that OFs variations affect calculated dose depending on the algorithm and on the delivery mode (field jaw/MLC-defined). The Monte Carlo (MC) algorithm resulted significantly more sensitive to OFs variations than all of the other algorithms. CONCLUSION: The aim of our study was to evaluate how small fields OFs inaccuracies can affect the dose calculation in VMAT brain radiosurgery treatments plans. It was observed that simulated OFs errors, return dosimetric calculation accuracies within the 3% between concurrent plans analysed in terms of percentage dose differences at selected dose/volume points of the PTV coverage and OARs sparing. ADVANCES IN KNOWLEDGE: First multicentre study involving different Planning/Linacs about undetectable errors in commissioning output factor for small fields.

Patient dose in angiographic interventional procedures: A multicentre study in Italy.

PURPOSE: The Council Directive 2013/59/EURATOM considers interventional radiology to be a special practice involving high doses of radiation and requiring strict monitoring to ensure the best quality assurance programs. This work reports the early experience of managing dose data from patients undergoing angiography in a multicentre study. MATERIALS AND METHODS: The study was based on a survey of about 15,200 sample procedures performed in 21 Italian hospitals centres involved on a voluntary basis. The survey concerned the collection of data related to different interventional radiology procedures: interventional cardiology, radiology, neuroradiology, vascular surgery, urology, endoscopy and pain therapy from a C-Arm and fixed units. The analysis included 11 types of procedures and for each procedure, air-kerma, kerma-area product and fluoroscopy time were collected. RESULTS: The duration and dose values of fluoroscopic exposure for each procedure is strongly dependent on individual clinical circumstances including the complexity of the procedure; the observed distribution of patient doses was very wide, even for a specified protocol. The median values of the parameters were compared with the diagnostic reference levels (DRL) proposed for some procedures in Italy (ISTISAN) or internationally. This work proposes local DRL values for three procedures. CONCLUSION: This first data collection serves to take stock of the situation on patient's dosimetry in several sectors and is the starting point for obtaining and updating DRL recalling that these levels are dependent on experience and technology available.

Novel 3D heart left ventricle muscle segmentation method for PET-gated protocol and its verification.

OBJECTIVE: The aim of this study was to propose and verify a universal method of left ventricular myocardium segmentation, able to operate on heart gated PET data with different sizes, shapes and uptake distributions. The proposed method can be classified as active model method and is based on the BEAS (B-spline Explicit Active Surface) algorithm published by Barbosa et al. The method was implemented within the Pmod PCARD software package. Method verification by comparison with reference software and phantom data is also presented in the paper. METHODS: The proposed method extends the BEAS model by defining mechanical features of the model: tensile strength and bending resistance. Formulas describing model internal energy increase during its stretching and bending are proposed. The segmentation model was applied to the data of 60 patients, who had undergone cardiac gated PET scanning. QGS by Cedars-Sinai and ECTb by Emory University Medical Centre served as reference software for comparing ventricular volumes. The method was also verified using data of left ventricular phantoms of known volume. RESULTS: The results of the proposed method are well correlated with the results of QGS (slope: 0.841, intercept: 0.944 ml, R2: 0.867) and ECTb (slope: 0.830, intercept: 2.109 ml, R2: 0.845). The volumes calculated by the proposed method were very close to the true cavity volumes of two different phantoms. CONCLUSIONS: The analysis of gated PET data by the proposed method results in volume measurements comparable to established methods. Phantom experiments demonstrate that the volume values correspond to the physical ones.

Lung stereotactic ablative body radiotherapy: A large scale multi-institutional planning comparison for interpreting results of multi-institutional studies.

PURPOSE: A large-scale multi-institutional planning comparison on lung cancer SABR is presented with the aim of investigating possible criticism in carrying out retrospective multicentre data analysis from a dosimetric perspective. METHODS: Five CT series were sent to the participants. The dose prescription to PTV was 54Gy in 3 fractions of 18Gy. The plans were compared in terms of PTV-gEUD2 (generalized Equivalent Uniform Dose equivalent to 2Gy), mean dose to PTV, Homogeneity Index (PTV-HI), Conformity Index (PTV-CI) and Gradient Index (PTV-GI). We calculated the maximum dose for each OAR (organ at risk) considered as well as the MLD2 (mean lung dose equivalent to 2Gy). The data were stratified according to expertise and technology. RESULTS: Twenty-six centers equipped with Linacs, 3DCRT (4% - 1 center), static IMRT (8% - 2 centers), VMAT (76% - 20 centers), CyberKnife (4% - 1 center), and Tomotherapy (8% - 2 centers) collaborated. Significant PTV-gEUD2 differences were observed (range: 105-161Gy); mean-PTV dose, PTV-HI, PTV-CI, and PTV-GI were, respectively, 56.8±3.4Gy, 14.2±10.1%, 0.70±0.15, and 4.9±1.9. Significant correlations for PTV-gEUD2 versus PTV-HI, and MLD2 versus PTV-GI, were observed. CONCLUSIONS: The differences in terms of PTV-gEUD2 may suggest the inclusion of PTV-gEUD2 calculation for retrospective data inter-comparison.

Is prostate-specific antigen percentage decrease predictive of clinical outcome after permanent iodine-125 interstitial brachytherapy for prostate cancer?

OBJECTIVE: To determine the usefulness of prostate-specific antigen (PSA) percentage (vs. pretreatment value assumed as 100%) in prediction of biochemical relapse, after iodine-125 ((125)I) permanent brachytherapy for prostate cancer, to employ a parameter independent by the initial PSA amount and by the individual prostatic volume. METHODS AND MATERIALS: Our study included 133 patients, 102 still disease free (Group A) and 31 who experienced proven biochemical recurrence (Group B). PSA levels before and after (125)I brachytherapy were recorded, and PSA percentage vs. pretreatment values were calculated. Cox regression model, receiver operating characteristic curves, and Kaplan-Meier regression model with log-rank test were calculated. RESULTS: We observed that, in patients submitted to brachytherapy for prostate cancer, a PSA percentage >20% of pretreatment value is highly associated with relapse risk (p<0.0001) and that this association is strongly present since t=6 months of followup (p<0.0001), with a hazard ratio near to five times (4.965), a sensitivity of 72.4%, and specificity of 79.8% related to the chosen cutoff. DISCUSSION: Despite the amount of PSA is the only parameter that the clinicians can deploy to monitor patient's followup after permanent interstitial brachytherapy for prostate cancer, its evolution in time seems unable to predict early biochemical relapse as it is influenced by prostatic volume and initial PSA amount. CONCLUSIONS: Our data suggest that a PSA percentage >20% of pretreatment value at 6 months might represent an early, inexpensive, and useful predictive tool of bad outcome in patients after permanent brachytherapy.