Automated segmentation and measurement of global white matter lesion volume in patients with multiple sclerosis

Alfano, Bruno; Brunetti, Arturo; Larobina, Michele; Quarantelli, Mario; Tedeschi, Enrico; Ciarmiello, Andrea; Covelli, Eugenio Maria; Salvatore, Marco

doi:10.1002/1522-2586(200012)12:6<799::aid-jmri2>3.0.co;2-#

Cited by 89 publications

(46 citation statements)

References 21 publications

(19 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other authors employed sensitivity (Sens), specificity (Spec) and accuracy (Acc) (Akselrod-Ballin et al, 2009;Alfano et al, 2000;Hadjiprocopis and Tofts, 2003;Liu et al, 2009;Sajja et al, 2006;TomasFernandez and Warfield, 2011;Wels et al, 2008;Wu et al, 2006): These measures should be considered carefully, as TP (lesions) are much smaller than TN (normal appearing brain tissues). An automatic segmentation in which the TLL is 10 times greater than the ground truth can still result in Sens = 100%, Spec = 99.3%, and Acc = 99.3% when the TLL is around 1 cm 3 and brain size is 1,500 cm 3 .…”

Section: -Overlap Measuressupporting

confidence: 63%

“…Depending on the algorithm employed, all features might also need to be normalized to have the same variance or zero mean. Some authors have proposed a segmentation method based on the normalization of images, which is in turn based on the intensity of the training datasets (Alfano et al, 2000;Erickson and Avula, 1998). For example, one normalization approach employed tissue that is constant across the images and not affected by the disease (i.e., subcutaneous fat) (Erickson and Avula, 1998).…”

Section: Supervised Learning Methodsmentioning

confidence: 99%

“…For example, one normalization approach employed tissue that is constant across the images and not affected by the disease (i.e., subcutaneous fat) (Erickson and Avula, 1998). Another approach used spin-echo sequences with specific parameters to generate T1, T2, and PD relaxometry maps (Alfano et al, 2000). Once all images have the same intensity range, tissues can be classified using fixed intensity rules for all patients in the normalized multidimensional intensity feature space.…”

Section: Supervised Learning Methodsmentioning

confidence: 99%

“…In addition, some metrics are sensitive to the surface-to-volume ratio of the structures segmented; therefore, it may be important to take lesion size into account to correctly understand the results. Only 14 papers included the lesion load of the patients, and the range varied greatly among the papers; the lowest lesion loads ranged from approximately 1 cm 3 (Alfano et al, 2000;García-Lorenzo et al, 2009;Harmouche et al, 2006) to 8 cm 3 (Shiee et al, 2009), and the highest, from 20.0 cm 3 to 130 cm 3 (Alfano et al, 2000). Some authors divided the patients according to lesion load (Khayati et al, 2008a;Sajja et al, 2006), although no consensus exists on how to achieve this division.…”

Section: Characteristics Of the Databasementioning

confidence: 99%

See 3 more Smart Citations

Review of automatic segmentation methods of multiple sclerosis white matter lesions on conventional magnetic resonance imaging

García-Lorenzo

Francis

Narayanan

et al. 2013

Medical Image Analysis

306

263

View full text Add to dashboard Cite

Magnetic resonance (MR) imaging is often used to characterize and quantify multiple sclerosis (MS) lesions in the brain and spinal cord. The number and volume of lesions have been used to evaluate MS disease burden, to track the progression of the disease and to evaluate the effect of new pharmaceuticals in clinical trials. Accurate identification of MS lesions in MR images is extremely difficult due to variability in lesion location, size and shape in addition to anatomical variability between subjects. Since manual segmentation requires expert knowledge, is time consuming and is subject to intra-and interexpert variability, many methods have been proposed to automatically segment lesions.The objective of this study was to carry out a systematic review of the literature to evaluate the state of the art in automated multiple sclerosis lesion segmentation. From 1240 hits found initially with PubMed and Google scholar, our selection criteria identified 80 papers that described an automatic lesion segmentation procedure applied to MS. Only 47 of these included quantitative validation with at least one realistic image. In this paper, we describe the complexity of lesion segmentation, classify the automatic MS lesion segmentation methods found, and review the validation methods applied in each of the papers reviewed. Although many segmentation solutions have been proposed, including some with promising results using MRI data obtained on small groups of patients, no single method is widely employed due to performance issues related to the high variability of MS lesion appearance and differences in image acquisition. The challenge remains to provide segmentation techniques that work in all cases regardless of the type of MS, duration of the disease, or MRI protocol, and this within a comprehensive, standardized validation framework. MS lesion segmentation remains an open problem.

show abstract

Section: -Overlap Measuressupporting

confidence: 63%

Section: Supervised Learning Methodsmentioning

confidence: 99%

Section: Supervised Learning Methodsmentioning

confidence: 99%

Section: Characteristics Of the Databasementioning

confidence: 99%

See 2 more Smart Citations

Review of automatic segmentation methods of multiple sclerosis white matter lesions on conventional magnetic resonance imaging

García-Lorenzo

Francis

Narayanan

et al. 2013

Medical Image Analysis

306

263

View full text Add to dashboard Cite

show abstract

“…Alfano et al [56] developed an automated approach based on relaxometric and geometric features for classification of MS lesions from 3-D MRI images. Boudraa et al [57] employed the FCM algorithm to 1.5 T twodimensional (2-D) MRI images for classifying normal and abnormal brain structures.…”

Section: Multiple Sclerosis Diagnosismentioning

confidence: 99%