BACKGROUND:Bone metastasis is a common event in lung cancer progression. Early diagnosis of lung malignant tumor with bone metastasis is crucial for selecting effective treatment strategies. However, 14.3% of patients are still difficult to diagnose after SPECT/CT examination.PURPOSE:Machine learning analysis of [99mTc]-methylene diphosphate (99mTc-MDP) SPECT/CT scans to distinguish bone metastases from benign bone lesions in patients with lung cancer.METHODS:One hundred forty-one patients (69 with bone metastases and 72 with benign bone lesions) were randomly assigned to the training group or testing group in a 7:3 ratio. Lesions were manually delineated using ITK-SNAP, and 944 radiomics features were extracted from SPECT and CT images. The least absolute shrinkage and selection operator (LASSO) regression was used to select the radiomics features in the training set, and the single/bimodal radiomics models were established based on support vector machine (SVM). To further optimize the model, the best bimodal radiomics features were combined with clinical features to establish an integrated Radiomics-clinical model. The diagnostic performance of models was evaluated using receiver operating characteristic (ROC) curve and confusion matrix, and performance differences between models were evaluated using the Delong test.RESULTS:The optimal radiomics model comprised of structural modality (CT) and metabolic modality (SPECT), with an area under curve (AUC) of 0.919 and 0.907 for the training and testing set, respectively. The integrated model, which combined SPECT, CT, and two clinical features, exhibited satisfactory differentiation in the training and testing set, with AUC of 0.939 and 0.925, respectively.CONCLUSIONS:The machine learning can effectively differentiate between bone metastases and benign bone lesions. The Radiomics-clinical integrated model demonstrated the best performance.

PurposeThis study aimed to analyze the recurrence patterns of thoracic esophageal squamous cell carcinoma (ESCC) after radical surgery, and to understand its implication in the clinical target volume (CTV) design of postoperative radiotherapy (PORT) in patients with ESCC.Methods and materialsA total of 428 recurrent ESCC patients after radical surgery between 2014 and 2018 were included in this study. Recurrence patterns, especially anastomotic and regional lymph node recurrence (LNR), were analyzed. A T-shaped CTV were proposed for PORT and were evaluated whether it could cover most of regional LNR.ResultsThese patients all experienced anastomotic and/or regional LNR. Among the 428 patients, 27 cases (6.3%) had anastomotic recurrence only, and184 cases (43.0%) had LNR only. Those sites with an LNR rate higher than 15% in upper thoracic ESCC were as follows: No.101, No.104R, No.104L, No.106recR, No.106recL, No.106pre, No.106tb, No.107, and No. 109. Those with middle thoracic ESCC were as follows: No.104R, No.104L, 106recR, No.106recL, No.106pre, No.106tb, and No.107. Lastly, individuals with lower thoracic ESCC were as follows: No.104L, 106recR, No.106recL, No. 106pre, No. 106tb, No.107, and abdominal No. 3. The proportion of LNR not included in the proposed T-shaped CTV was 12.5% (1/8), 4.7% (6/128), and 10.4% (5/48) in the upper, middle, and lower thoracic segments, respectively.ConclusionsLNR was the most common type of local-regional recurrence in patients after radical surgery. Supraclavicular, superior and middle mediastinal lymph nodes had the highest recurrence rate, the rate of LNR which was outside T-shaped PORT CTV we proposed was less than 15%.

Purpose: The aim of this study was to investigate risk factors for esophageal squamous cell carcinoma (ESCC) treated with radiotherapy (RT) with or without chemotherapy to guide how to reduce the occurrence of esophageal fistula (EF). Methods: 414 patients with ESCC who underwent RT with or without chemotherapy were collected in Shandong Cancer Hospital from February 2012 to June 2018 retrospectively. The clinical characters and dosimetric parameters were recorded. Univariate and multivariate logistic regression analyses were provided to determine the risk factors associated with EF. Results: The cumulative incidences of EF were 11.1% (46/414 patients). The median follow-up period was 15.8 months (range: 0.2-82.6months). The median survival time (MST) of patients with EF was 5.3 months. In univariate analysis, age, T4 stage, N3 stage, chemotherapy regimens, re-RT, ulcerative esophageal cancer (EC), esophageal stenosis, the maximum thickness of the tumor and the length of tumor had a correlation with the prevalence of EF. In multivariable logistic regression analysis, T4 stage, N3 stage, re-RT, ulcerative EC, esophageal stenosis, the maximum thickness of the tumor was confirmed as risk factors for EF. Conclusion: This study revealed that T4 stage, N3 stage, re-RT, ulcerative EC, esophageal stenosis, the maximum thickness of the tumor were risk factors associated with EF. We ought to attach importance to the prevention of EF. Patients with risk factors for EF should be paid close attention.

Objective To investigate the feasibility of online and offline cone-beam CT (CBCT) guided radiotherapy for lung cancer. Methods Fourteen patients with lung tumor treated by three-dimen-sional conformal radiotherapy were investigated. Online kV CBCT scan,image registration and setup correc-tion were performed before and immediately after radiotherapy. CBCT online-guided correction data were used to calculate the population-based CTV-PTV margins under the condition of non-correction and correction in every fraction respectively. The numbers of initial images and the population-based CTV-PTV margins af-ter the offline compensation of the system setup error were evaluated with the permission of 0.5 mm and 1.5 mm maximal residue error,respectively. Results Under the condition of non-correction,the required mar-gins for total error were 5.7 mm,8.0 mm and 7.8 mm in the left-right(x axis) ,cranio-caudal(y axis) and anterior-posterior(z axis) directions, respectively. When the tumor was corrected in every fraction, the re-quired margins for intra-fraction error were 2.4 mm,2.4 mm and 2.3 mm in x,y and z axes, respectively. To correct the systematic setup error,9 sets of CBCT images for 3.3 mm,3.7 mm and 3.6 mm PTV margins, and 7 sets of CBCT images for 3.9 mm,4.3 mm and 4.3 mm PTV margins in x,y and z axes were necessary when 0. 5 mm and 1.5 mm maximal residue errosr were permited respectively. Conclusions Both of the online CBCT correction and the offline adaptive correction can markedly reduce the impact of setup error and reduce the required PTV margins accordingly. It is feasible to deliver the online and offline image guided ra-diation for patients with lung tumor.