Search

Purpose: The proposed work aims to develop an algorithm to precisely segment the lung parenchyma in thoracic CT scans. To achieve this goal, the proposed technique utilized a combination of deep learning and traditional image processing algorithms. The initial step utilized a trained convolutional neural network (CNN) to generate preliminary lung masks, followed by the proposed post-processing algorithm for lung boundary correction. Methods: First, the proposed method trained an improved 2D U-Net CNN model with Inception-ResNet-v2as its backbone. The model was trained on 32 CT scans from two different sources: one from the VESSEL12 grand challenge and the other from AIIMS Delhi. Further, the model’s performance was evaluated on a test dataset of 16 CT scans with juxta-pleural nodules obtained from AIIMS Delhi and the LUNA16 challenge. The model’s performance was assessed using evaluation metrics such as average volumetric dice coefficient (DSCavg), average IoU score (IoUavg), and average F1 score (F1avg). Finally, the proposed post-processing algorithm was implemented to eliminate false positives from the model's prediction and to include juxta-pleural nodules in the final lung masks. Results: The trained model reported a DSCavgof 0.9791 ± 0.008, IoUavgof 0.9624 ± 0.007, and F1avgof 0.9792 ± 0.004 on the test dataset. Applying the post-processing algorithm to the predicted lung masks obtained a DSCavgof 0.9713 ± 0.007, IoUavgof 0.9486 ± 0.007, and F1avgof 0.9701 ± 0.008. The post-processing algorithm successfully included juxta-pleural nodules in the final lung mask. Conclusions: Using a CNN model, the proposed method for lung parenchyma segmentation produced precise segmentation results. Furthermore, the post-processing algorithm addressed false positives and negatives in the model’s predictions. Overall, the proposed approach demonstrated promising results for lung parenchyma segmentation. The method has the potential to be valuable in the advancement of computer-aided diagnosis (CAD) systems for automatic nodule detection.

Purpose: The current study explores the application of 3D U-Net architectures combined with Inception and ResNet modules for precise lung nodule detection through deep learning-based segmentation technique. This investigation is motivated by the objective of developing a Computer-Aided Diagnosis (CAD) system for effective diagnosis and prognostication of lung nodules in clinical settings. Methods: The proposed method trained four different 3D U-Net models on the retrospective dataset obtained from AIIMS Delhi. To augment the training dataset, affine transformations and intensity transforms were utilized. Preprocessing steps included CT scan voxel resampling, intensity normalization, and lung parenchyma segmentation. Model optimization utilized a hybrid loss function that combined Dice Loss and Focal Loss. The model performance of all four 3D U-Nets was evaluated patient-wise using dice coefficient and Jaccard coefficient, then averaged to obtain the average volumetric dice coefficient (DSCavg) and average Jaccard coefficient (IoUavg) on a test dataset comprising 53 CT scans. Additionally, an ensemble approach (Model-V) was utilized featuring 3D U-Net (Model-I), ResNet (Model-II), and Inception (Model-III) 3D U-Net architectures, combined with two distinct patch sizes for further investigation. Results: The ensemble of models obtained the highest DSCavgof 0.84 ± 0.05 and IoUavgof 0.74 ± 0.06 on the test dataset, compared against individual models. It mitigated false positives, overestimations, and underestimations observed in individual U-Net models. Moreover, the ensemble of models reduced average false positives per scan in the test dataset (1.57 nodules/scan) compared to individual models (2.69–3.39 nodules/scan). Conclusions: The suggested ensemble approach presents a strong and effective strategy for automatically detecting and delineating lung nodules, potentially aiding CAD systems in clinical settings. This approach could assist radiologists in laborious and meticulous lung nodule detection tasks in CT scans, improving lung cancer diagnosis and treatment planning.

ObjectivesMore than 50% patients develop emesis during induction therapy for acute myeloid leukaemia (AML). The addition of aprepitant for emesis control in children receiving induction for AML have not been explored.MethodsA single-institutional randomised, open-label trial (NCT02979548) was conducted where children between 5 and 18 years with the diagnosis of AML being planned for 3+7 induction regimen were included. All study participants received ondansetron (0.15 mg/kg) every 8 hours for 8 days starting 30 min prior to chemotherapy. Children belonging to aprepitant group additionally received aprepitant capsules (15–40 kg=days 1–3, 80 mg; >40 kg=day 1, 125 mg and days 2–3, 80 mg) starting from 1 hour prior to chemotherapy. The proportion of patients with complete response (CR) in chemotherapy induced vomiting (CIV) in acute phase (day 1–8), delayed phase (day 9–13), overall and initial 96 hours were recorded along with severity of vomiting and adverse effects.ResultsTotal 78 children were randomised (Aprepitant group: 37 and control group: 41). The proportion of patients with CR in CIV was significantly higher in Aprepitant group in acute phase (p=0.007), overall phase (p=0.007) and in initial 96 hours (p<0.001) but it was not different in delayed phase (p=0.07). The first episode of vomiting was also significantly delayed in aprepitant group (p=0.02). Adverse effect profile was similar in two groups.ConclusionAprepitant significantly improves emesis control in children receiving induction therapy for AML, especially in acute phase and should be routinely incorporated as part of antiemetic prophylaxis.Trial registration numberThe study was registered at ClinicalTrials.gov (NCT02979548).

ObjectivesTo assess the cost-effectiveness of addition of olanzapine to a prophylactic antiemetic regimen containing aprepitant, dexamethasone and ondansetron among children receiving highly emetogenic chemotherapy (HEC) in India, Bangladesh, Indonesia, the UK and the USA.MethodsHealth states were estimated using individual patient-level outcome data from a randomised trial. The incremental cost–utility ratio (ICUR), incremental cost-effectiveness ratio and net monetary benefit (NMB) were calculated from the patient perspective for India, Bangladesh, Indonesia, the UK and the USA. One-way sensitivity analysis was done by varying the cost of olanzapine, cost of hospitalisation and utility values by ±25%.ResultsThe olanzapine arm had an increment of 0.0018 quality-adjusted life-years (QALY) over the control arm. The mean total expenditure in the olanzapine arm was greater by US$0.51, US$0.43, US$6.73, US$11.05 and US$12.35 in India, Bangladesh, Indonesia, the UK and the USA, respectively. The ICUR($/QALY) was US$282.60 in India, US$241.42 in Bangladesh, US$3755.93 in Indonesia, US$6161.83 in the UK and US$6887.41 in the USA. The NMB was US$9.86, US$10.12, US$14.08, US$44.74 and US$98.79 for India, Bangladesh, Indonesia, the UK and the USA, respectively. The ICUR estimates of the base case and sensitivity analysis were below the willingness-to-pay threshold in all scenarios.ConclusionThe addition of olanzapine as a fourth agent for antiemetic prophylaxis is cost-effective despite an increase in overall expenditure. Olanzapine should be uniformly considered for children receiving HEC.

Kikuchi-Fujimoto disease is a rare, benign, and self-limiting condition that mostly affects young females. Cervical lymphadenopathy with fever is the most common presentation of the disease. It may have unusual presentations that can lead to diagnostic dilemma and delay in diagnosis. We report a case of a 25-year-old female who presented with relapsing fever and cervical lymphadenopathy. Because of atypical presentation, there was a delay in diagnosis and increase in morbidity. High index of suspicion with collaboration between clinicians and pathologists is essential for early and accurate diagnosis of the disease.