Machine learning classification of mediastinal lymph node metastasis in NSCLC: a multicentre study in a Western European patient population.

BACKGROUND: [18F] FDG PET-CT has an important role in the initial staging of lung cancer; however, accurate differentiation between activity in malignant and benign intrathoracic lymph nodes on PET-CT scans can be challenging. The purpose of the current study was to investigate the effect of incorporating primary tumour data and clinical features to differentiate between [18F] FDG-avid malignant and benign intrathoracic lymph nodes. METHODS: We retrospectively selected lung cancer patients who underwent PET-CT for initial staging in two centres in the Netherlands. The primary tumour and suspected lymph node metastases were annotated and cross-referenced with pathology results. Lymph nodes were classified as malignant or benign. From the image data, we extracted radiomic features and trained the classifier model using the extreme gradient boost (XGB) algorithm. Various scenarios were defined by selecting different combinations of data input and clinical features. Data from centre 1 were used for training and validation of the models using the XGB algorithm. To determine the performance of the model in a different hospital, the XGB model was tested using data from centre 2. RESULTS: Adding primary tumour data resulted in a significant gain in the performance of the trained classifier model. Adding the clinical information about distant metastases did not lead to significant improvement. The performance of the model in the test set (centre 2) was slightly but statistically significantly lower than in the validation set (centre 1). CONCLUSIONS: Using the XGB algorithm potentially leads to an improved model for the classification of intrathoracic lymph nodes. The inclusion of primary tumour data improved the performance of the model, while additional knowledge of distant metastases did not. In patients in whom metastases are limited to lymph nodes in the thorax, this may reduce costly and invasive procedures such as endobronchial ultrasound or mediastinoscopy procedures.

Measuring the actual I-131 thyroid uptake curve with a collar detector system: a feasibility study.

Radionuclide therapy using I-131 is commonly used for the treatment of benign thyroid diseases. The therapeutic dose to be administered is calculated based on the type of disease, the volume of the thyroid, and the measured uptake percentage. This methodology assumes a similar biological half-life of iodine, whereas in reality a large variation in biological half-life is observed. More knowledge about the actual biological half-life of iodine for individual patients will improve the quantification of the delivered radiation dose during radioiodine therapy and could aid the evaluation of the success of the therapy. In this feasibility study we used a novel measurement device [Collar Therapy Indicator (CoTI)] to measure the uptake curve of patients undergoing I-131 radioiodine therapy. The CoTI device is a light-weight wearable device that contains two independent gamma radiation detectors that are placed in a collar. By comparing results of thyroid uptake measurements with results obtained with a gamma camera, the precision of the system is demonstrated. Additionally, for three patients the uptake curve is measured during 48 h of admission in the hospital. The presented results demonstrate the feasibility of the new measurement device to measure the uptake curve during radioiodine therapy.

Contrast-noise-ratio (CNR) analysis and optimisation of breast-specific gamma imaging (BSGI) acquisition protocols.

BACKGROUND: Breast cancer is one of the most prevalent forms of cancer in women. Breast-specific gamma imaging (BSGI) is a diagnostic imaging method that uses sestamibi-labelled 99Tc and a dedicated gamma camera to localize malignant lesions in breast tissue. The aim of this study is to investigate if the current acquisition protocol for BSGI at our hospital is optimized for the detection of lesions in our patients. METHODS: We analyzed patient data and performed a phantom study with a Dilon 6800 gamma camera. The patient data were collected from a group of 13 patients (740 MBq 99mTc-sestamibi, four views per patient were dynamically acquired with a frame duration of 30 s per frame and a total acquisition time of 8 min per view). Reduced-time static images were created, and contrast-to-noise ratios of identified hotspots were determined for different acquisition times. For the phantom study, we used a contrast detail phantom to investigate the contrast and resolution properties, within the range of relevant clinical acquisition parameters. The phantom was filled with a concentration of 80 MBq in 500 ml of water, and we dynamically acquired frames for a total acquisition time of 60 min using a general purpose (GP) collimator. To compare the GP collimator with the high-resolution collimator, a second acquisition was made for both collimators with a total acquisition time of 16 min. RESULTS: The initial analysis of BSGI scans of the 13 patients showed that a dose reduction by a factor of 3 would not have reduced the number of observable hotspots in each of the acquired views. However, a subsequent systematic analysis of our protocol with a contrast-detail phantom showed that dose reduction results in a lower observability of hotspots, whereas increased doses resulted in a higher observability. CONCLUSION: We believe that the results of our phantom study are relevant for clinical practice and that further dose reduction cannot be recommended for the BSGI exams at our hospital and that an increase of the administered activity should be considered.