MRI-based classification of the anatomical variants of the hippocampal head.

PURPOSE: The head of the hippocampus (H) is classically described as having two to four digitations both in ex vivo specimens and in vivo MR coronal images. The aim of this study was to develop and evaluate a new MR-based classification of the anatomical variants of the hippocampal head in a large sample population of healthy subjects. METHODS: MR images of the brain of 238 young healthy subjects (138 men and 100 women; age range 18-39) were analyzed. The head of the H was identified on coronal reformatted 3D T1 weighted MR images. The frequencies were reported for hemisphere and sex. Inter-rater reliability was assessed. RESULTS: Eight variants of the hippocampal head were described. Class 0 (11.4%) indicated a total absence of sulci. This class was further subdivided as follows: 0A (one digitation, 10.1%) and 0B (no digitations or "null variant", 1.3%). Class 1 (25.6%) presented a single sulcus and was further subdivided into four types according to the location and the width of the sulcus [1A (8.8%), 1B (12.8%), 1C (1.3%), and 1D (2.7%)]. Class 2 (63.0%, the most frequent and the classical variant) had two symmetrical sulci and three digitations. Statistically significant differences between the two hemispheres were observed only in women and overall. Differences in prevalence between sexes were not observed. CONCLUSIONS: The large study population allowed the description of a novel morphological classification of the different anatomical variants of normal H in the coronal plane. This classification could reduce the risk of misinterpreting normal anatomical variants as pathological.

Impact of the IASLC/ATS/ERS classification in pN0 pulmonary adenocarcinomas: a study with radiological-pathological comparisons and survival analyses.

The aim of this study was: (1) to compare the new pathological findings as detected by the IASLC/ATS/ERS classification with the traditional radiological features in pulmonary pN0 adenocarcinomas, (2) to evaluate their prognostic significance on overall survival (OS). A total of 42 surgically resected pN0 pulmonary adenocarcinomas were analyzed. On CT scans, the following radiological data were recorded: sphericity, predominant margins, cavitation and bronchogram, attenuation and percentage of ground glass opacity (GGO). On pathological examination, tumors were categorized according to the IASLC/ATS/ERS classification; Sica score and grade, pathological stage, tumor major axis, pleural invasion, vascular and lymphatic invasion, peritumoral lymphoid infiltration, and cytological features were also determined. Clinical follow up was available in 37 cases (range 1-117 months). Radiologically, 31 solid and 11 semisolid tumors were found. Morphologically, 2 minimally invasive and 40 invasive adenocarcinomas were diagnosed. In radiological-pathological comparisons, (1) the acinar pattern was higher in tumors with solid attenuation and low GGO (p=0.018); (2) the lepidic pattern was more elevated in tumors with high GGO (p=0.012). In multivariate survival analyses with stage, predominant margins on CT scans (p=0.036) and Sica score (p=0.028) significantly affected OS. This study confirms the validity of the new classification of pulmonary adenocarcinomas in radiological-pathological comparisons and underlines the importance of both radiological and pathological findings in correctly identifying their prognostic features.

Intra- and interoperator variability of lobar pulmonary volumes and emphysema scores in patients with chronic obstructive pulmonary disease and emphysema: comparison of manual and semi-automated segmentation techniques.

PURPOSE: We aimed to compare the intra- and interoperator variability of lobar volumetry and emphysema scores obtained by semi-automated and manual segmentation techniques in lung emphysema patients. MATERIALS AND METHODS: In two sessions held three months apart, two operators performed lobar volumetry of unenhanced chest computed tomography examinations of 47 consecutive patients with chronic obstructive pulmonary disease and lung emphysema. Both operators used the manual and semi-automated segmentation techniques. The intra- and interoperator variability of the volumes and emphysema scores obtained by semi-automated segmentation was compared with the variability obtained by manual segmentation of the five pulmonary lobes. RESULTS: The intra- and interoperator variability of the lobar volumes decreased when using semi-automated lobe segmentation (coefficients of repeatability for the first operator: right upper lobe, 147 vs. 96.3; right middle lobe, 137.7 vs. 73.4; right lower lobe, 89.2 vs. 42.4; left upper lobe, 262.2 vs. 54.8; and left lower lobe, 260.5 vs. 56.5; coefficients of repeatability for the second operator: right upper lobe, 61.4 vs. 48.1; right middle lobe, 56 vs. 46.4; right lower lobe, 26.9 vs. 16.7; left upper lobe, 61.4 vs. 27; and left lower lobe, 63.6 vs. 27.5; coefficients of reproducibility in the interoperator analysis: right upper lobe, 191.3 vs. 102.9; right middle lobe, 219.8 vs. 126.5; right lower lobe, 122.6 vs. 90.1; left upper lobe, 166.9 vs. 68.7; and left lower lobe, 168.7 vs. 71.6). The coefficients of repeatability and reproducibility of emphysema scores also decreased when using semi-automated segmentation and had ranges that varied depending on the target lobe and selected threshold of emphysema. CONCLUSION: Semi-automated segmentation reduces the intra- and interoperator variability of lobar volumetry and provides a more objective tool than manual technique for quantifying lung volumes and severity of emphysema.