Early prediction of acute xerostomia during radiation therapy for nasopharyngeal cancer based on delta radiomics from CT images

Abstract

Background: Acute xerostomia is the most common side effect of radiation therapy (RT) for head and neck (H&N) malignancies. Investigating radiation-induced changes of computed tomography (CT) radiomics in parotid glands (PGs) and saliva amount (SA) can predict acute xerostomia during the RT for nasopharyngeal cancer (NPC).Methods: CT and SA data from 35 patients with stages I-IVB were randomly collected from an NPC clinical trial registered on the clinicaltrials.gov (ID: NCT01762514). All patients received radical treatment based on intensity-modulated RT (IMRT) with a prescription dose of 68.1 Gy in 30 fractions. The patients' ages ranged 24-72 years, and each patient had five CT sets acquired at treatment position: at the 0 th , 10 th , 20 th , 30 th fractions during the RT, and at 3-month later after the RT. The PGs for each CT set were delineated by a radiation oncologist and verified independently by another. Patients' saliva was collected every other 10 days during the RT. Acute xerostomia was evaluated based on the RTOG acute toxicity scoring and the SA. In total, 1,703 radiomics features were calculated for PGs from each CT set, including feature value at 0 th fraction (FV 0F ), FV 10F , and delta FV (ΔFV 10F-0F ), respectively. Extensive experiments were conducted to achieve the optimal results. RidgeCV and Recursive Feature Elimination (RFE) were used for feature selection, while linear regression was used for predicting SA 30F . Four more patients were added for independent testing.Results: Substantial changes in various radiomics metrics of PGs were observed during the RT. Eight normalized feature value (NFV), selected from NFV 0F , predicted SA 10F with a mean square error (MSE) of 0.9042 and a R 2 score of 0.7406. Fourteen NFV, selected from ΔNFV 10F-0F , NFV 0F , and NFV 10F to predict SA 30F , showed the best predictive ability with an MSE of 0.0569. The model predicted the level of acute xerostomia with a precision of 0.9220 and a sensitivity of 100%, compared to the clinical observed SA. For the independent test, the MSE of PSA 30F was 0.0233. Conclusions:This study demonstrated that radiation-induced acute xerostomia level could be early predicted based on the SA and radiomics changes of the PGs during IMRT delivery. SA, NFV 0F , NFV 10F , and especially ΔNFV 10F-0F provided the best performance on acute xerostomia prediction for individual patient based on RidgeCV_RFE_LinearRegression method of delta radiomics.

Application of Radiomics for the Prediction of Radiation-Induced Toxicity in the IMRT Era: Current State-of-the-Art

“…In this perspective, the identification of quantitative imaging parameters correlated with both acute and late xerostomia is of paramount importance. Changes over time of radiomics features (delta-radiomics) have been extensively evaluated both in terms of acute and late xerostomia prediction (35)(36)(37)(38). In an effort to better elucidate the relationship between parotid gland shrinkage after RT and late xerostomia, van Dijk et al (37) recently demonstrated a correlation between delta radiomics surface changes in contralateral parotid gland and late xerostomia in 68 patients (AUC 0.93 in test cohort).…”

Longitudinal acquisition repeatability of MRI radiomics features: An ACR MRI phantom study on two MRI scanners using a 3D T1W TSE sequence

“…Longitudinal stability of MRI radiomics feature should be of great importance for some clinical applications, such as MRI‐guided radiotherapy (MRIgRT or MRgRT), 29–31 using either integrated MRI and linear accelerator systems (MR‐LINAC) 32–34 for on‐line guidance or dedicated MR‐simulators (MR‐sims) for off‐line guidance 35–37 . The longitudinal daily MRI data generated in MRgRT might hold potentials for close monitoring and evaluation of treatment response, and thus treatment adaptation based on not only morphological/function changes of target tumors but also the interfractional change of radiomics features during the treatment course, which is known as delta‐radiomics 38,39 . Longitudinal radiomics feature instability, if cannot be clearly differentiated from true treatment‐induced tumor change, might lead to compromised or even false‐positive response evaluation and thus influence clinical decision.…”

Application of radiomics and machine learning in head and neck cancers

“…Acute xerostomia is the most common side effect of radiation therapy for HNC. Pota et al and Liu et al [78][79] conducted a study on HNC and NPC, respectively. They obtained CT scans of patients before, during, and after treatment to obtain imaging features and establish the best model to use during the initial treatment phase to predict the development of acute xerostomia after radiation therapy in cancer patients.…”