Current students

Radiation therapy (RT) represents an option for treating lung, lymphoma, esophagus, and breast cancer. Recent studies have demonstrated that the dose of radiation to the heart limits the two-year survival for the cohort of lung cancer patients. Specifically, volumes of the heart receiving a dose between 5 Gy and 30 Gy are more strongly associated with a higher risk of death.The study will be conducted on a population of 700 stage III lung cancer patients treated between 2013 and 2021 in various Italian and European centers. The dataset will include information on patient characteristics, tumor details, treatment indications, two-year survival, and quantitative information from DICOM RT (Digital Imaging and Communications in Medicine Radiation Therapy) data. DICOM RT represents the standardized form of collecting imaging data. In the case of RT, the file includes planning CT slices, contoured organs, and the dose plan distribution. In summary, the file allows extraction of patient-specific parameters related to imaging and dose.
The study's first goal will be extracting dosimetric features for Heart Substructures (HSs) using AI contour software. This will involve the delineation of the main interest area, such as the aorta, the Left and Right Atriums, the Left and Right Ventricles, the Inferior Vena Cava, the Superior Vena Cava, and the Pulmonary Artery. By contouring these new structures, it will be possible to define the relative dosimetric variables. From the CT scan, we will derive morphological information and quantitative score to assess the presence of calcification within the heart. It has been discovered that calcifications are associated with a higher risk of developing heart dysfunctions.
The second objective of the study is to model 2-year Overal Survival (OS2) after RT for LA-NSCLS patients using the predictive model already introduced by Nix [1]. This model includes three components: the Equivalent dose (EQD2) to the tumor, the chemotherapy impact on the metastatic process, and the Survival Limiting Toxicity (SLT), which refers to reduced survival at higher doses of cancer. Due to the limited availability of data in the meta-analysis by Nix and colleagues, they consider SLT as a mathematical function of the dose to the tumor, representing a weak surrogate of the dose to healthy tissues.
As a novelty of the project, we improved the SLT function by:
replacing the dose to the tumor with a set of significant dose parameters coming from the HSs distribution;
adding the dose to the healthy lung (lung tissue - tumoral target);
introducing the cardiac calcification score.
Moreover, we incremented the description of the factors increasing the OS2 taking into account the parameters describing the immunotherapy (not yet implemented in the clinic at the time of [1]).
The weight of the new relevant factors will be estimated through the maximum-likelihood method. The new model will be compared to the previous one using the likelihood ratio and the Akaike Information Criteria (AIC).
Finally, the derived set of dosimetric constraints for safely sparing the heart substructures will guide an in-silico comparison of different treatment modalities. Indeed, when a treatment plan does not satisfy these limits, the clinician could address the patient to other RT techniques more capable of satisfying the dose distribution requirements, testing the clinical gain derived from the use of Tomotherapy, hadron-therapy and MRI-Linac. The differences in the dose distribution will be converted through the model (specifically to the part for SLT) to estimate the risk reduction of major adverse cardiac events caused by such more conforming treatment modalities.