Inter- and intra-observer agreement in dermatologists' diagnoses of hyperpigmented facial lesions and development of an algorithm for automated diagnosis.

BACKGROUND: Hyperpigmentation has varied aetio-pathologies. Hence, accurate and reproducible diagnosis of the type of hyperpigmentation is essential for effective management. It is typically made clinically by dermatologists but the rate of inter- and intra-observer agreement/variability is unknown. Hyperpigmented facial lesions are extremely common but access to dermatological services is difficult or costly in most countries. Thus, it is desired to evaluate dermatologists' inter- and intra-observer agreement in the diagnosis and to develop an algorithm for automated diagnosis. MATERIALS AND METHODS: Hyperpigmented lesions on 392 facial images were diagnosed by three experienced dermatologists either jointly or independently, and this process was subsequently repeated for 52 randomly selected images. When there was non-concordance amongst the dermatologists for the diagnosis, a majority decision was taken as correct diagnosis. Inter-observer and intra-observer agreement were analysed for the diagnosis of the hyperpigmented lesions. Thereafter, a multiclass classification method was developed to perform the task in an automatic manner. The developed algorithm was compared and validated against the ground truth derived from the dermatologists. RESULTS: Both inter- and intra-observer agreements are in the moderate range. The algorithm agreed well with the derived ground truth, with a Kappa value of 0.492, which is similar to the Kappa values of inter- and intra- observer agreements. CONCLUSION: The rates of inter- and intra-observer agreement in the diagnosis of hyperpigmented facial lesions amongst dermatologists were moderate in this study. Compared to visual assessment from the dermatologists, automated diagnosis using the developed algorithm achieved a high rate of concordance.

Region-based snake with edge constraint for segmentation of lymph nodes on CT images.

Lymph nodes segmentation is a tedious process with large inter-user variability when performed manually. To facilitate lymph nodes assessment for lung cancer patient, we present an automatic and improved snake segmentation method for thoracic lymph nodes on CT images in this paper. We first investigated the performance of both edge-based and region-based snake algorithms for the segmentation task, using a B-spline contour parameterization. The effect of the number of B-spline control points on the snake performance was also examined. Both edge-based and region-based snakes were found to have their own advantages and disadvantages for lymph nodes segmentation. We further developed a method of region-based snake with edge constraint, which utilizes a self-adjusting mechanism to integrate both edge and region information in a constructive manner. The average Dice Similarity Coefficient obtained was 0.853 ± 0.059 and 0.841 ± 0.108 for the baseline and follow-up lymph nodes respectively using the proposed method. The method was found to be an effective lymph node segmentation method and would potentially be useful to help with treatment response evaluations in the clinical practice.

Automating the tracking of lymph nodes in follow-up studies of thoracic CT images.

The study of lymph node features over time is of great clinical significance. Tracking of the same lymph node in CT images over time is done manually in the current clinical practice, which is tedious and lack of consistency. In this paper, we propose a search scheme to automate the process. Regions of interest (ROIs) are located by mapping the center point of lymph node based on the transformation found in the rigid registration. Similarity values between ROI of the template image and ROIs of repository images are compared, the highest of which decides the best match. Our method generated a success rate of 82% in determining the corresponding image in follow-up scan with the same lymph node as in baseline. The location of the lymph node in the corresponding image is tracked and estimated by mapping the lymph node center at baseline image using the transformation obtained from both affine and free-form deformation (FFD) registration. FFD performs better than affine registration in tracking the lymph node location. All lymph nodes in our study are tracked successfully by the suggested points which fall within the boundary of the same node in the corresponding follow-up images using FFD registration.

Investigation on the use of deformable registrations in lymph node tracking over time.

Evolutionary features of lymph nodes such as change in size over time are decisive descriptors to assess disease progression in cancer patient. Other than study at one point in time, it is more useful to derive temporal analysis on structures of interest. The paper presents the use of deformable registration in lymph node tracking, particularly in the context of disease progression. We found that the extent of disease progression plays an important role in determining the performance of deformable registration in aligning up small anatomic structures, such as lymph node. Both Demons and B-spline registrations have their own advantages in different medical context.