In recent years, 2D convolutional neural networks (CNNs) have been extensively used to diagnose neurological diseases from magnetic resonance imaging (MRI) data due to their potential to discern subtle and intricate patterns. Despite the high performances reported in numerous studies, developing CNN models with good generalization abilities is still a challenging task due to possible data leakage introduced during cross-validation (CV). In this study, we quantitatively assessed the effect of a data leakage caused by 3D MRI data splitting based on a 2D slice-level using three 2D CNN models to classify patients with Alzheimer’s disease (AD) and Parkinson’s disease (PD). Our experiments showed that slice-level CV erroneously boosted the average slice level accuracy on the test set by 30% on Open Access Series of Imaging Studies (OASIS), 29% on Alzheimer’s Disease Neuroimaging Initiative (ADNI), 48% on Parkinson’s Progression Markers Initiative (PPMI) and 55% on a local de-novo PD Versilia dataset. Further tests on a randomly labeled OASIS-derived dataset produced about 96% of (erroneous) accuracy (slice-level split) and 50% accuracy (subject-level split), as expected from a randomized experiment. Overall, the extent of the effect of an erroneous slice-based CV is severe, especially for small datasets.

Lung densitometry assesses with computed tomography (CT) the X-ray attenuation of the pulmonary tissue which reflects both the degree of inflation and the structural lung abnormalities implying decreased attenuation, as in emphysema and cystic diseases, or increased attenuation, as in fibrosis. Five reasons justify replacement with lung densitometry of semi-quantitative visual scales used to measure extent and severity of diffuse lung diseases: (I) improved reproducibility; (II) complete vs. discrete assessment of the lung tissue; (III) shorter computation times; (IV) better correlation with pathology quantification of pulmonary emphysema; (V) better or equal correlation with pulmonary function tests (PFT). Commercially and open platform software are available for lung densitometry. It requires attention to technical and methodological issues including CT scanner calibration, radiation dose, and selection of thickness and filter to be applied to sections reconstructed from whole-lung CT acquisition. Critical is also the lung volume reached by the subject at scanning that can be measured in post-processing and represent valuable information per se. The measurements of lung density include mean and standard deviation, relative area (RA) at −970, −960 or −950 Hounsfield units (HU) and 1st and 15th percentile for emphysema in inspiratory scans, and RA at −856 HU for air trapping in expiratory scans. Kurtosis and skewness are used for evaluating pulmonary fibrosis in inspiratory scans. The main indication for lung densitometry is assessment of emphysema component in the single patient with chronic obstructive pulmonary diseases (COPD). Additional emerging applications include the evaluation of air trapping in COPD patients and in subjects at risk of emphysema and the staging in patients with lymphangioleiomyomatosis (LAM) and with pulmonary fibrosis. It has also been applied to assess prevalence of smoking-related emphysema and to monitor progression of smoking-related emphysema, alpha1 antitrypsin deficiency emphysema, and pulmonary fibrosis. Finally, it is recommended as end-point in pharmacological trials of emphysema and lung fibrosis.

One of the most important problems in the segmentation of lung nodules in CT imaging arises from possible attachments occurring between nodules and other lung structures, such as vessels or pleura. In this report, we address the problem of vessels attachments by proposing an automated correction method applied to an initial rough segmentation of the lung nodule. The method is based on a local shape analysis of the initial segmentation making use of 3-D geodesic distance map representations. The correction method has the advantage that it locally refines the nodule segmentation along recognized vessel attachments only, without modifying the nodule boundary elsewhere. The method was tested using a simple initial rough segmentation, obtained by a fixed image thresholding. The validation of the complete segmentation algorithm was carried out on small lung nodules, identified in the ITALUNG screening trial and on small nodules of the lung image database consortium (LIDC) dataset. In fully automated mode, 217/256 (84.8%) lung nodules of ITALUNG and 139/157 (88.5%) individual marks of lung nodules of LIDC were correctly outlined and an excellent reproducibility was also observed. By using an additional interactive mode, based on a controlled manual interaction, 233/256 (91.0%) lung nodules of ITALUNG and 144/157 (91.7%) individual marks of lung nodules of LIDC were overall correctly segmented. The proposed correction method could also be usefully applied to any existent nodule segmentation algorithm for improving the segmentation quality of juxta-vascular nodules.

Computed tomography (CT) is the most sensitive imaging technique for detecting lung nodules, and is now being evaluated as a screening tool for lung cancer in several large samples studies all over the world. In this report, we describe a semiautomatic method for 3-D segmentation of lung nodules in CT images for subsequent volume assessment. The distinguishing features of our algorithm are the following. 1) The user interaction process. It allows the introduction of the knowledge of the expert in a simple and reproducible manner. 2) The adoption of the geodesic distance in a multithreshold image representation. It allows the definition of a fusion--segregation process based on both gray-level similarity and objects shape. The algorithm was validated on low-dose CT scans of small nodule phantoms (mean diameter 5.3-11 mm) and in vivo lung nodules (mean diameter 5--9.8 mm) detected in the Italung-CT screening program for lung cancer. A further test on small lung nodules of Lung Image Database Consortium (LIDC) first data set was also performed. We observed a RMS error less than 6.6% in phantoms, and the correct outlining of the nodule contour was obtained in 82/95 lung nodules of Italung-CT and in 10/12 lung nodules of LIDC first data set. The achieved results support the use of the proposed algorithm for volume measurements of lung nodules examined with low-dose CT scanning technique.