PET is a popular medical imaging modality for various clinical applications, including diagnosis and image-guided radiation therapy. The low-dose PET (LDPET) at a minimized radiation dosage is highly desirable in clinic since PET imaging involves ionizing radiation, and raises concerns about the risk of radiation exposure. However, the reduced dose of radioactive tracers could impact the image quality and clinical diagnosis. In this paper, a supervised deep learning approach with a generative adversarial network (GAN) and the cycle-consistency loss, Wasserstein distance loss, and an additional supervised learning loss, named as S-CycleGAN, is proposed to establish a non-linear end-to-end mapping model, and used to recover LDPET brain images. The proposed model, and two recently-published deep learning methods (RED-CNN and 3D-cGAN) were applied to 10% and 30% dose of 10 testing datasets, and a series of simulation datasets embedded lesions with different activities, sizes, and shapes. Besides vision comparisons, six measures including the NRMSE, SSIM, PSNR, LPIPS, SUV max and SUV mean were evaluated for 10 testing datasets and 45 simulated datasets. Our S-CycleGAN approach had comparable SSIM and PSNR, slightly higher noise but a better perception score and preserving image details, much better SUV mean and SUV max , as compared to RED-CNN and 3D-cGAN. Quantitative and qualitative evaluations indicate the proposed approach is accurate, efficient and robust as compared to other stateof-the-art deep learning methods.

Introduction:Primary hepatic mucosa-associated lymphoid tissue (MALT) lymphoma is extremely rare and we herein report a case of a patient suffering from primary hepatic MALT lymphoma with concomitant hepatitis B virus infection.Diagnostic modalities and outcome:Double masses were found in a 59-year-old Chinese female patient. We reported the laboratory results, computed tomography (CT) and fluorine-18-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/CT images among other findings. As far as we know, only 9 cases have been reported till now using 18F-FDG PET/CT imaging. Our patient's lesions were found to conform to standard uptake values of FDG.Conclusion:It indicates that hepatic MALT lymphoma can be studied with 18F-FDG PET/CT like other 18F-FDG-avid lymphomas. It was also noted that delayed-time-point FDG PET imaging may further improve the detection of the MALT lymphoma in liver. Although the patient in this case refused further treatment, potential management options, including rituximab, which is also discussed in this review.

BackgroundTo evaluate the factors affecting the maximum standardized uptake value (SUVmax) of metastatic lymph nodes in different histological types of non-small cell lung cancer (NSCLC) on integrated positron emission tomography and computed tomography (PET-CT).MethodsThis was a retrospective, single-institution review of 122 patients with pathologically proven NSCLC who had PET-CT scanning at the same center. Lymph node metastases were pathologically confirmed in tissue specimens from surgical patients. Statistical evaluation of PET-CT results was performed on a per-nodal-station basis.ResultsThe tumor SUVmax of squamous cell carcinoma (SCC) (11.0 ± 4.1) was higher than that of adenocarcinoma (AC) (7.4 ± 4.4) (P < 0.01), however, the SUVmax of the metastatic lymph nodes did not differ between the SCC (4.6 ± 3.1) and AC groups(3.6 ± 2.5) (P = 0.221). The SUVmax of metastatic lymph nodes was positively correlated with lymph node size but not with the primary tumor SUVmax, primary tumor size, tumor location and tumor differentiation. The frequency of a SUVmax of lymph nodes ≥2.5 was 44%, 80%,100% in SCC group and 39%, 59%, 90% in AC group when the short-axis diameter of metastatic lymph node was <10 mm, 10–15 mm, and > 15 mm, respectively. The low sensitivity for metastatic lymph nodes on PET-CT was increased when the SUVmax cut-off for malignancy was considered to be above the normal background compared with that when the SUVmax cut-off was above 2.5.ConclusionsThere was no difference in the SUVmax of metastatic lymph nodes in the SCC and AC groups. The SUVmax of metastatic lymph nodes was positively correlated with metastatic lymph node size. There was a high false negative rate if lymph nodes with a short-axis diameter less than 10 mm and a extremely low false negative rate if lymph nodes with a short-axis diameter higher than 15 mm. Although an increased sensitivity may be achieved by decreasing the SUVmax cut-off, invasive staging may still be required for negative lymph nodes due to the lower sensitivity of PET-CT in both SCC and AC.

Background: This study classifies lung adenocarcinoma (ADC) and squamous cell carcinoma (SCC) using subregion-based radiomics features extracted from positron emission tomography/computed tomography (PET/CT) images.

PurposeTo propose and evaluate habitat imaging-based 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) radiomics for preoperatively discriminating non-small cell lung cancer (NSCLC) and benign inflammatory diseases (BIDs).MethodsThree hundred seventeen 18F-FDG PET/CT scans were acquired from patients who underwent aspiration biopsy or surgical resection. All volumes of interest (VOIs) were semiautomatically segmented. Each VOI was separated into variant subregions, namely, habitat imaging, based on our adapted clustering-based habitat generation method. Radiomics features were extracted from these subregions. Three feature selection methods and six classifiers were applied to construct the habitat imaging-based radiomics models for fivefold cross-validation. The radiomics models whose features extracted by conventional habitat-based methods and nonhabitat method were also constructed. For comparison, the performances were evaluated in the validation set in terms of the area under the receiver operating characteristic curve (AUC). Pairwise t-test was applied to test the significant improvement between the adapted habitat-based method and the conventional methods.ResultsA total of 1,858 radiomics features were extracted. After feature selection, habitat imaging-based 18F-FDG PET/CT radiomics models were constructed. The AUC of the adapted clustering-based habitat radiomics was 0.7270 ± 0.0147, which showed significantly improved discrimination performance compared to the conventional methods (p &lt;.001). Furthermore, the combination of features extracted by our adaptive habitat imaging-based method and non-habitat method showed the best performance than the other combinations.ConclusionHabitat imaging-based 18F-FDG PET/CT radiomics shows potential as a biomarker for discriminating NSCLC and BIDs, which indicates that the microenvironmental variations in NSCLC and BID can be captured by PET/CT.