Commissioning of a novel PET-Linac for biology-guided radiotherapy (BgRT).

BACKGROUND: Biology-guided radiotherapy (BgRT) is a novel radiotherapy delivery technique that utilizes the tumor itself to guide dynamic delivery of treatment dose to the tumor. The RefleXion X1 system is the first radiotherapy system developed to deliver SCINTIX® BgRT. The X1 is characterized by its split arc design, employing two 90-degree positron emission tomography (PET) arcs to guide therapeutic radiation beams in real time, currently cleared by FDA to treat bone and lung tumors. PURPOSE: This study aims to comprehensively evaluate the capabilities of the SCINTIX radiotherapy delivery system by evaluating its sensitivity to changes in PET contrast, its adaptability in the context of patient motion, and its performance across a spectrum of prescription doses. METHODS: A series of experimental scenarios, both static and dynamic, were designed to assess the SCINTIX BgRT system's performance, including an end-to-end test. These experiments involved a range of factors, including changes in PET contrast, motion, and prescription doses. Measurements were performed using a custom-made ArcCHECK insert which included a 2.2 cm spherical target and a c-shape structure that can be filled with a PET tracer with varying concentrations. Sinusoidal and cosine4 motion patterns, simulating patient breathing, was used to test the SCINTIX system's ability to deliver BgRT during motion-induced challenges. Each experiment was evaluated against specific metrics, including Activity Concentration (AC), Normalized Target Signal (NTS), and Biology Tracking Zone (BTZ) bounded dose-volume histogram (bDVH) pass rates. The accuracy of the delivered BgRT doses on ArcCHECK and EBT-XD film were evaluated using gamma 3%/2 mm and 3%/3 mm analysis. RESULTS: In static scenarios, the X1 system consistently demonstrated precision and robustness in SCINTIX dose delivery. The end-to-end delivery to the spherical target yielded good results, with AC and NTS values surpassing the critical thresholds of 5 kBq/mL and 2, respectively. Furthermore, bDVH analysis consistently confirmed 100% pass rates. These results were reaffirmed in scenarios involving changes in PET contrast, emphasizing the system's ability to adapt to varying PET avidities. Gamma analysis with 3%/2 mm (10% dose threshold) criteria consistently achieved pass rates > 91.5% for the static tests. In dynamic SCINTIX delivery scenarios, the X1 system exhibited adaptability under conditions of motion. Sinusoidal and cosine4 motion patterns resulted in 3%/3 mm gamma pass rates > 87%. Moreover, the comparison with gated stereotactic body radiotherapy (SBRT) delivery on a conventional c-arm Linac resulted in 93.9% gamma pass rates and used as comparison to evaluate the interplay effect. The 1 cm step shift tests showed low overall gamma pass rates of 60.3% in ArcCHECK measurements, while the doses in the PTV agreed with the plan with 99.9% for 3%/3 mm measured with film. CONCLUSIONS: The comprehensive evaluation of the X1 radiotherapy delivery system for SCINTIX BgRT demonstrated good agreement for the static tests. The system consistently achieved critical metrics and delivered the BgRT doses per plan. The motion tests demonstrated its ability to co-localize the dose where the PET signal is and deliver acceptable BgRT dose distributions.

Adaptive Region-Specific Loss for Improved Medical Image Segmentation.

Defining the loss function is an important part of neural network design and critically determines the success of deep learning modeling. A significant shortcoming of the conventional loss functions is that they weight all regions in the input image volume equally, despite the fact that the system is known to be heterogeneous (i.e., some regions can achieve high prediction performance more easily than others). Here, we introduce a region-specific loss to lift the implicit assumption of homogeneous weighting for better learning. We divide the entire volume into multiple sub-regions, each with an individualized loss constructed for optimal local performance. Effectively, this scheme imposes higher weightings on the sub-regions that are more difficult to segment, and vice versa. Furthermore, the regional false positive and false negative errors are computed for each input image during a training step and the regional penalty is adjusted accordingly to enhance the overall accuracy of the prediction. Using different public and in-house medical image datasets, we demonstrate that the proposed regionally adaptive loss paradigm outperforms conventional methods in the multi-organ segmentations, without any modification to the neural network architecture or additional data preparation.

Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met) Utilizing Free-Text Clinical Narratives.

We propose a deep learning model - Probabilistic Prognostic Estimates of Survival in Metastatic Cancer Patients (PPES-Met) for estimating short-term life expectancy (>3 months) of the patients by analyzing free-text clinical notes in the electronic medical record, while maintaining the temporal visit sequence. In a single framework, we integrated semantic data mapping and neural embedding technique to produce a text processing method that extracts relevant information from heterogeneous types of clinical notes in an unsupervised manner, and we designed a recurrent neural network to model the temporal dependency of the patient visits. The model was trained on a large dataset (10,293 patients) and validated on a separated dataset (1818 patients). Our method achieved an area under the ROC curve (AUC) of 0.89. To provide explain-ability, we developed an interactive graphical tool that may improve physician understanding of the basis for the model's predictions. The high accuracy and explain-ability of the PPES-Met model may enable our model to be used as a decision support tool to personalize metastatic cancer treatment and provide valuable assistance to the physicians.