Classification Of Dysplasia Research Articles

Evaluation of trochlear dysplasia using MRI: correlation between the classification system of Dejour and objective parameters of trochlear dysplasia

Trochlear dysplasia is considered to be one of the major factors causing patellofemoral instability (PFI). Dejour's classification is widely used to assess the severity of trochlear dysplasia. Additionally, in current literature, different quantitative parameters are recommended to distinguish between a normal trochlea and a dysplastic trochlea. In order to achieve a more objective evaluation of the trochlea, the aim of this study was to evaluate whether specific measurements of the femoral trochlea can be assigned to the qualitative classification system of Dejour. Transverse MRI T2-weighted scans of 80 knees with symptomatic PFI and varying severity of trochlear dysplasia were classified according to Dejour (type A to D). For all MRI scans, quantitative measurements with parameters as described in the literature were applied. The values were then allocated to Dejour's classification. In addition to the four-grade analysis, two-grade analysis was also performed (Dejour type A against type BCD). Dependent on the cut-off values, specificity, sensitivity and Youden index for each parameter was defined. The allocation resulted in the following distribution: type A trochlear dysplasia n = 25, type B n = 23, type C n = 18 and type D n = 14. In descriptive statistics, none of the measurements proposed in the literature could be assigned to the four-grade classification system of Dejour. For the two-grade analysis at the cut-off, sensitivity ranged from 75 to 86 % and specificity from 76 to 84 % for lateral trochlear inclination, trochlear facet asymmetry and depth of trochlear groove. All other measurements showed a poor sensitivity ranging from 49 to 67 % and specificity from 40 to 72 %. Interobserver and intraobserver repeatability for the measured parameters was fair to moderate (ICC values 0.34-0.58) in high-grade dysplasia (type BCD) and substantial to almost perfect (ICC values 0.71-0.88) in low-grade trochlear dysplasia (type A). Quantitative measurements of the femoral trochlea have shown to be of limited value for the assessment of trochlear dysplasia. None of the quantitative measurements of the trochlea on transverse images could be assigned to the four-grade descriptive classification of trochlear dysplasia of Dejour. Additionally, measurements could not be reliably performed in high-grade trochlear dysplasia. However, trochlear inclination, trochlear facet asymmetry and depth of trochlear groove may help to distinguish between low-grade and high-grade dysplasia.

Classifications of glenoid dysplasia, glenoid bone loss and glenoid loosening: a review of the literature

So far, glenoid implantation still remains a challenge and is technically demanding even for an experienced shoulder surgeon. Each shoulder pathology has its own evolution. In primary glenohumeral osteoarthritis, glenoid involvement and proper morphology vary considerably. Erosion is more posterior and inferior. In rheumatoid arthritis, glenoid erosion is more medial with a very weak and soft bone. In eccentric arthritis, glenoid erosion is most of the time superior. Glenoid component loosening has been recognized as one of the common indications for revision surgery after total shoulder arthroplasty. Scapular notching is specific to the reverse shoulder arthroplasty. Moreover, there is concern about the high frequency of glenoid components that demonstrate radiographic periprosthetic lucencies. There is little information available to guide clinical decision making regarding glenoid surgery. Placement or replacement with a standard glenoid component is usually possible. In some instances, bone graft reconstruction or the use of augmented or custom components can be an option. The purpose of this study is to evaluate the main glenoid erosion classifications.

Focal Preauricular Dermal Dysplasia in a Newborn

We report a case of focal preauricular dermal dysplasia in an 18-day-old healthy girl. We discuss the classification of focal preauricular dermal dysplasia within the spectrum of focal facial dermal dysplasia and aplasia cutis congenita.

Hippocampal sclerosis—Origins and imaging

The association between hippocampal sclerosis (HS) and epilepsy has been known for almost two centuries. For many years, HS was studied in postmortem series; however, since the mid-20th century, surgical specimens from temporal lobe resections have provided important new knowledge. HS is the most common pathology underlying drug-resistant mesial temporal lobe epilepsy (MTLE), a syndrome with a characteristic history and seizure semiology. In the early 1990 s, it was recognized that magnetic resonance imaging (MRI) could detect HS. The standard MRI protocol for temporal lobe abnormalities uses coronal slices perpendicular to the long axis of the hippocampus. The MRI features of HS include reduced hippocampal volume, increased signal intensity on T(2)-weighted imaging, and disturbed internal architecture. The histopathologic diagnosis of HS is usually straightforward, with neuronal loss and chronic fibrillary gliosis centered on the pyramidal cell layer. There are several patterns or subtypes of HS recognized from surgical series based on qualitative or quantified assessments of regional neuronal loss. The pathologic changes of HS include granule cell dispersion, mossy fiber sprouting, and alterations to interneurons. There may also be more extensive sclerosis of adjacent structures in the medial temporal lobe, including the amygdala and parahippocampal gyrus. Subtle cortical neuropathologies may accompany HS. The revised classification of dysplasias in epilepsy denotes these as focal cortical dysplasias type IIIa. Sometimes, HS occurs with a second lesion, either in the temporal lobe or extratemporal, most often ipsilateral to the HS. HS on preoperative MRI strongly predicts good seizure outcome following temporal lobe resection (TLR). If adequate MRI shows no structural correlate in patients with MTLE, functional imaging studies are valuable, especially if they are in agreement with ictal electroencephalography (EEG) findings. Focal hypometabolism on 18F-fluorodeoxyglucose-positron emission tomography (FDG-PET) ipsilateral to the symptomatic temporal lobe predicts a good surgical outcome; the added value of (11) C-Flumazenil-PET (FMZ-PET) and proton magnetic resonance spectroscopy (MRS) is less clear. Surgical methods have evolved, particularly resecting less tissue, aiming to preserve function without compromising seizure outcome. Around two thirds of patients operated for MTLE with HS obtain seizure freedom. However, the best surgical approach to optimize seizure outcome remains controversial.

Radiographic parameters associated with lateral patella degeneration in young patients

Localized articular cartilage degeneration in the patellofemoral joint is a common but yet understudied condition in younger patients. The purpose of this paper was to determine whether there are significant differences in radiographic alignment between those with and without isolated lateral patellofemoral degeneration. Subjects with isolated symptomatic lateral patellofemoral degeneration and control subjects with no radiographic degeneration participated in the study. Variables of interest included the Caton-Deschamps index, sulcus angle, lateral patellofemoral angle, Dejour classification of trochlear dysplasia, patella linear displacement and the tibial tubercle-trochlear groove distance. We found significant differences between the patellofemoral degeneration group versus control group, respectively, for the Caton-Deschamps index (1.12 ± 0.1 vs. 1.00 ± 0.1), lateral patellofemoral angle (10.6 ± 4.3 vs. 16.6 ± 5.5) and tibial tubercle-trochlear groove distance (16.6 ± 4.0 vs. 9.0 ± 4.3). However, we found no difference in the sulcus angle (141.2 ± 8.2 vs. 137.0 ± 6.0), patella linear displacement (3.7 ± 1.9 vs. 4.0 ± 1.7) or in the Dejour Classification. It appears that isolated lateral patellofemoral degeneration is associated with specific radiographic indices. Even though the radiographic measures in patients with degeneration may not be considered pathologic, they are "high normal" and may represent a risk factor for the development of focal chondral degeneration in the lateral trochlea and patella. III.

Diagnosis and management of fibromuscular dysplasia: an expert consensus

Diagnosis and management of fibromuscular dysplasia: an expert consensus

Ki-67 and ProExC are useful immunohistochemical markers in esophageal squamous intraepithelial neoplasia

Esophageal squamous intraepithelial neoplasia has been widely recognized as a precursor lesion for esophageal squamous cell carcinoma. Early detection offers the best prognosis for esophageal squamous cell carcinoma. The differentiation of squamous dysplasia from reactive change and the classification of squamous dysplasia into high-grade or low-grade are sometimes subjective and challenging. In this study, we sought to evaluate multiple biomarkers and to develop clinically useful adjunct tools for difficult esophageal squamous intraepithelial neoplasia cases. Immunohistochemical stains using antibodies against Ki-67, ProExC, p16, and p53 were performed on esophageal biopsy or resection specimens from 25 patients including 35 foci of high-grade dysplasia and 25 foci of low-grade dysplasia, and from 10 control cases containing 52 foci of normal/reactive hyperplasia. In situ hybridization tests for human papillomavirus were performed in 11 cases. The immunostains for all 4 markers were scored as negative, intermediate, and strong according to established criteria. Intermediate and strong Ki-67 and ProExC staining showed similar detecting power and exhibited very high sensitivity and specificity for distinguishing normal/reactive hyperplasia from esophageal squamous intraepithelial neoplasia and normal/reactive hyperplasia from low-grade esophageal squamous intraepithelial neoplasia. Strong Ki-67 staining was exclusively seen in high-grade esophageal squamous intraepithelial neoplasia, which provided additional value in distinguishing high-grade from low-grade esophageal squamous intraepithelial neoplasia. Strong ProExC staining was also seen in most high-grade esophageal squamous intraepithelial neoplasia foci (80%). Although the frequencies of intermediate/strong staining patterns of p53 increased with increasing degree of dysplasia, the sensitivity of p53 was much lower than that of Ki-67 and ProExC. p16 did not show consistent immunostain pattern in the normal/reactive hyperplasia and esophageal squamous intraepithelial neoplasia. Two (18%) of 11 tested cases were positive for human papillomavirus infection. This study demonstrates that both Ki-67 and ProExC can be used as an adjunct tool for diagnosing difficult cases of esophageal squamous intraepithelial neoplasia.

Classification of trochlear dysplasia as predictor of clinical outcome after trochleoplasty

Sulcus-deepening trochleoplasty restores the trochlear groove in patients with patellofemoral instability and underlying trochlear dysplasia. There are types of dysplasia both with (B and D) and without (A and C) a supratrochlear spur. The aim of this study was to identify influencing factors for the clinical outcome following trochleoplasty. Forty-four knees in 38 patients who underwent trochleoplasty for instability (type A in 9, B in 15, C in 9 and D in 11 knees) were assessed clinically with the Kujala score and radiologically with radiographs and MRI. The median follow-up was 4 (2-7.8) years. At follow-up, the median Kujala score had improved from 68 (29-84) to 90 (42-100) points (P<0.001). Instability (P<0.001) and pain (P=0.027) decreased significantly, but in 3 knees, pain was worse postoperatively. Twenty-seven knees were ranked as excellent, 10 as good, 2 as fair and 5 as poor. Overall, dysplasia types B and D benefited more from surgery than types A and C. The postoperative MRI revealed no chondrolysis or subchondral necrosis, but deterioration of cartilage on the lateral trochlear facet was identified. Trochleoplasty is a useful and reliable surgical technique to improve patellofemoral instability in patients with a dysplastic trochlea. While improved stability is predictable, pain is less predictable and may even increase following surgery. The overall results were directly dependent on the type of the dysplasia, with a significantly better clinical outcome in type B and D. The clinical relevance of this study is that severe dysplasia can successfully be treated with trochleoplasty. III.

Barrett’s oesophagus: past, present and future

Barrett's oesophagus (BE) is the most common cause of oesophageal adenocarcinoma. Adenocarcinoma develops through a meta-plasia-dysplasia-carcinoma sequence, but there are many unknown and controversial issues related to the diagnosis, classification, pathological criteria, molecular progression, natural history and treatment of metaplastic and dysplastic precursor lesions in BE. The purpose of this lecture (review) is to provide an evidence-based summary of the diagnosis and classification of BE-associated metaplasia and dysplasia, to provide a summary of the key pathological features of the most common types of precursor lesions, to outline common areas of diagnostic difficulties, and to evaluate adjunctive markers of cancer progression in BE. The current American College of Gastroenterologists (ACG) definition of BE is partially flawed, since not all cases of BE are endoscopically recognisable, non-goblet epithelium is biologically intestinalised and genetically abnormal, and determination of the presence or absence of goblet cells is susceptible to sampling error and variability with time. Metaplastic epithelium in BE shows clonal molecular abnormalities that precede morphological expression of dysplasia. Dysplasia in BE begins in the crypt bases and then extends more superficially to include the upper portion of the crypts and surface epithelium with time and progression. Low and high grade dysplasia are distinguished by the presence of marked cytological and/or architectural abnormalities in the latter compared to the former, but there are newly discovered subtypes of dysplasia, such as the foveolar and serrated type, in which differentiation of low from high grade dysplasia is difficult and the significance of grading dysplasia is unknown. For instance, one type of dysplasia (foveolar dysplasia) may occur in non-goblet metaplastic columnar epithelium and shows a high progression rate to carcinoma compared to conventional (intestinal type) dysplasia. There are few, if any, reliable adjunctive diagnostic techniques that can help differentiate non-dysplastic from dysplastic epithelium in BE, but α-methylacyl coenzyme A racemase (AMACR) staining has been shown to be useful in several different studies. Both low and high grade dysplasia are progressive lesions, and in general, the extent of dysplasia is a strong risk factor for progression to carcinoma. Of all the biological and genetic biomarkers studied to date, evaluation of DNA content is the most reliable and specific marker of progression in BE. The management of patients with dysplasia is variable between institutions, and ranges from aggressive surveillance, endoscopic mucosal resection, either with or without mucosal ablation, or total oesophagectomy. Recent data on radiofrequency ablation in BE is promising.

Expression of Phosphorylated Histone H2AX (γ-H2AX) in Normal and Neoplastic Squamous Epithelia of the Uterine Cervix

The histone γ-H2AX is a marker of activated DNA damage and is overexpressed in different cancers and their precursor lesions, indicating a role in oncogenic transformation. Epidermal growth factor receptor (EGFR) is overexpressed in many kinds of epithelial neoplasms. This study aimed to determine whether immunohistochemical expression of γ-H2AX is involved in the progression of the morphological spectrum from normal squamous cervical epithelia (NE, n=33) to cervical intraepithelial neoplasia (CIN; CIN1, n=9; CIN2/3, n=33) and cervical invasive squamous cell carcinoma (ISCC, n=33), whether γ-H2AX expression follows the pattern of proliferation of atypical squamous cells as shown by EGFR immunoreactivity, and whether it is correlated with clinicopathologic variables in ISCC. Immunostaining for both the factors was scored semiquantitatively for moderate and strong intensities. Gamma-H2AX and EGFR expression, respectively, increased from NE and CIN1 to CIN2/3 and ISCCs significantly (P=0.0001, respectively). Gamma-H2AX reactivity was found in the nuclei of the cells of the upper epithelial levels and the cells of basal/parabasal type in variable quantities in NE and CIN; expression of γ-H2AX was seen in the nuclei of ISCC including keratinizing cells in horn pearls. EGFR staining was mainly membranous and noted in basal/parabasal cells in NE and atypically proliferating keratinocytes in CIN and nonkeratinizing cells of ISCC. In addition, immature squamous metaplasias were decorated by the antibodies used. Immunoscores for γ-H2AX and EGFR, respectively, did not differ between International Federation of Gynecology and Obstetrics stages I and II, keratinizing and nonkeratinizing ISCC, and CIN2/3 and ISCC. However, expression patterns were different for both the factors, suggesting their involvement in different biological mechanisms, with regard to γ-H2AX apart from proliferation. Overexpression of γ-H2AX in CIN2/3 and ISCC of the uterine cervix reflects the neoplastic transformation of cervical squamous epithelia in reaction to increased DNA-damage and supports the classification of dysplasia into low-grade squamous intraepithelial lesions and high-grade squamous intraepithelial lesions.

Image analysis for classification of dysplasia in Barrett’s esophagus using endoscopic optical coherence tomography

Barrett’s esophagus (BE) and associated adenocarcinoma have emerged as a major health care problem. Endoscopic optical coherence tomography is a microscopic sub-surface imaging technology that has been shown to differentiate tissue layers of the gastrointestinal wall and identify dysplasia in the mucosa, and is proposed as a surveillance tool to aid in management of BE. In this work a computer-aided diagnosis (CAD) system has been demonstrated for classification of dysplasia in Barrett’s esophagus using EOCT. The system is composed of four modules: region of interest segmentation, dysplasia-related image feature extraction, feature selection, and site classification and validation. Multiple feature extraction and classification methods were evaluated and the process of developing the CAD system is described in detail. Use of multiple EOCT images to classify a single site was also investigated. A total of 96 EOCT image-biopsy pairs (63 non-dysplastic, 26 low-grade and 7 high-grade dysplastic biopsy sites) from a previously described clinical study were analyzed using the CAD system, yielding an accuracy of 84% for classification of non-dysplastic vs. dysplastic BE tissue. The results motivate continued development of CAD to potentially enable EOCT surveillance of large surface areas of Barrett’s mucosa to identify dysplasia.

Comparison of the clinical significance of the Papanicolaou test interpretations LSIL cannot rule out HSIL and ASC‐H

Despite the two-tiered classification of dysplasia in The Bethesda System (TBS), rare cases fall into the category squamous intraepithelial lesion (SIL) of indeterminate grade. These Pap tests are often interpreted as "LSIL/ASC-H" or "LSIL" with a comment indicating the presence of cells with features approaching HSIL. Patients with LSIL/ASC-H have a significant risk of CIN 2 or worse (29-61.5%) on follow-up cervical biopsies, similar to the risk of CIN 2 or worse in patients with ASC-H Pap tests (24-68%). The purpose of this study was to compare patients with ASC-H and LSIL/ASC-H Pap tests. Women with LSIL/ASC-H had a slightly lower incidence of CIN 2 or worse (PPV = 35.6%, 95% CI: 29.8-41.4%) on follow-up cervical biopsy than the control ASC-H group (PPV = 40.2%, 95% CI: 31.9-56.3%); this difference was not statistically significant. The difference in the distribution of the biopsy results between the two groups was statistically significant (P < 0.001). The current guidelines for the management of cervical cytologic abnormalities from the American Society for Colposcopy and Cervical Pathology (ASCCP) advocate similar treatment algorithms for both LSIL and ASC-H. The main difference is the option of cytologic follow-up or HPV testing for certain "special populations," as an alternative to colposcopy, for LSIL Pap test results. Based on our results, we recommend (1)LSIL/ASC-H to be added to TBS classification and (2) Pap test cases of LSIL/ASC-H may need to be clinically followed in a manner similar to ASC-H, i.e., colposcopy for all patients.

Serrated adenomas

Serrated adenomas are categorized as sessile serrated adenomas (SSAs) and traditional serrated adenomas (TSAs). SSAs are more prevalent in the proximal colon and lack classic dysplasia, whereas TSAs are more prevalent in the rectosigmoid and have cytologic dysplasia. Serrated adenomas may progress to colorectal adenocarcinoma through diverse molecular alterations. Colonoscopy is the only test for the early detection of serrated adenomas that allows inspection of the entire colon and same-session biopsy sampling or polypectomy, if necessary. If an endoscopic biopsy at the right colon reveals SSA without cytologic dysplasia or biopsy at the rectosigmoid reveals SSA or TSA, those polyps should be excised or surgically resected as necessary. Postpolypectomy surveillance for removed SSAs without dysplasia and TSAs must be performed at 5- and 3-year intervals, respectively, with colonoscopy to prevent recurrence and progression to colorectal adenocarcinoma.

Perspective on the classification of ectodermal dysplasia

The need for revisiting the classification of ectodermal dysplasia syndromes has come. Prior to considering a new way to classify this group of disorders, however, thought should be given to some basic ideas about terminology and the process of classification. Consequently, this article reiterates the meanings of the words ectoderm, dysplasia, genetic and syndrome, and describes the process by which numeric taxonomists go about classification. In the process, a new family with the Jorgenson syndrome is described and the inheritance of the Schopf syndrome is clarified. Finally, the backgrounds of those with vested interests in the ectodermal dysplasia syndromes are described in order that all approaches to classification are covered.

Rationale and background as basis for a new classification of the ectodermal dysplasias

Ectodermal dysplasias are heterogeneous heritable conditions characterized by congenital defects of one or more ectodermal structures and their appendages. Of approximately 200 different ectodermal dysplasias, about 30 have been identified at molecular level with identification of the causative gene. Itin and Fistarol emphasized that rather commonly non-fully expressed phenotypes exist, which make a clinical diagnosis more difficult. Freire-Maia and Pinheiro used the clinical aspects for their classification and Priolo integrated molecular genetic and clinical aspects for her scheme. Those two more historical classification schemes have the difficulty that when applied strictly, several additional groups of disorders should be integrated within the term of ectodermal dysplasias, for example, keratodermas with skin or hair alterations or the ichthyoses with associated abnormalities. Such consequent classification would lead to an endless list of conditions and would be useless for practical work. Recent evidence implicates a genetic defect in different pathways orchestrating ectodermal organogenesis. Modern molecular genetics will increasingly elucidate the basic defects of the different syndromes and yield more insight into the regulatory mechanisms of morphogenesis. In this way a reclassification of ectodermal dysplasias will be possible according to the function of their involved mutated genes. I will focus on the fact that with molecular methods it is possible to diagnose oligosymptomatic forms of ectodermal dysplasia. This is much more common than earlier anticipated and with the classification of ectodermal dysplasia on the basis of molecular diagnosis a new avenue is opened for symptom complexes which were impossible to classify in former times.

Precision, reproducibility, and clinical usefulness of measuring the Norberg angle by means of computerized image analysis

To evaluate the precision, reproducibility, and clinical usefulness of measuring the Norberg angle (NA) by means of a computerized system of image analysis. 1,182 consecutive radiographs of hip joints of various breeds of dogs assessed for hip dysplasia and 72 radiographs of hip joints of German Shepherd Dogs. Radiographs were assessed by a panel of 4 experts in consensus, and NAs were measured by means of a computerized system. Results of classification of hip dysplasia according to the Fédération Cynologique Internationale (FCI) and NAs were compared within dogs and among breeds. Measurement of the NA by means of image analysis was twice as reproducible as that via calipers. Mean NA of left hip joints was 0.38 degrees higher than that of right hip joints. The NA values accurately discriminated between hip joints of dogs without or with hip dysplasia, provided the values were also expressed as percentile rank based on the cumulative frequency distribution of NAs within the breed, and had good power to discriminate among various FCI classifications of hip joints. Mean NA for each dog breed as calculated by use of the lower of 2 NAs for each dog was highly variable and was moderately correlated with the existence of hip dysplasia (r = 0.5). Computer-assisted measurement of the NA was useful in assessing hip joint quality and can be implemented for quality control and standardization of the FCI classification and for international comparisons.

Histological assessment of oesophageal columnar mucosa

This review presents the pathological features of Barrett's oesophagus, with an emphasis on the role of pathologists in the diagnosis, surveillance and treatment of the disease. The diagnosis of Barrett's oesophagus is based both on endoscopy and histology. The surveillance of patients relies on systematic biopsy sampling, looking for dysplasia - intraepithelial neoplasia. Well established classifications of dysplasia are now used by pathologists, but there remain problems with this marker. Therefore, many alternative biomarkers have been proposed, that remain of limited interest in daily practice, including DNA-ploidy, proliferation markers, and p53 abnormalities. Endoscopic improvements already allow a better selection of biopsies, and it may be that new technologies will allow 'virtual biopsies'. The role of pathologists is now extended to the evaluation of new therapeutic modalities of early neoplastic lesions in Barrett's oesophagus, especially endoscopic mucosal resection.

Oral epithelial dysplasia classification systems: predictive value, utility, weaknesses and scope for improvement

At a workshop coordinated by the WHO Collaborating Centre for Oral Cancer and Precancer in the United Kingdom issues related to potentially malignant disorders of the oral cavity were discussed by an expert group. The consensus views of the Working Group are presented in a series of papers. In this report, we review the oral epithelial dysplasia classification systems. The three classification schemes [oral epithelial dysplasia scoring system, squamous intraepithelial neoplasia and Ljubljana classification] were presented and the Working Group recommended epithelial dysplasia grading for routine use. Although most oral pathologists possibly recognize and accept the criteria for grading epithelial dysplasia, firstly based on architectural features and then of cytology, there is great variability in their interpretation of the presence, degree and significance of the individual criteria. Several studies have shown great interexaminer and intraexaminer variability in the assessment of the presence or absence and the grade of oral epithelial dysplasia. The Working Group considered the two class classification (no/questionable/ mild - low risk; moderate or severe - implying high risk) and was of the view that reducing the number of choices from 3 to 2 may increase the likelihood of agreement between pathologists. The utility of this need to be tested in future studies. The variables that are likely to affect oral epithelial dysplasia scoring were discussed and are outlined here; these need to be researched in longitudinal studies to explore the biological significance of a low-risk or high-risk dysplasia.

A study on the early detection of colon cancer using the methods of wavelet feature extraction and SVM classifications of FTIR

This paper introduces a new method for the early detection of colon cancer using a combination of feature extraction based on wavelets for Fourier Transform Infrared Spectroscopy (FTIR) and classification using the Support Vector Machine (SVM). The FTIR data collected from 36 normal SD rats, 60 1,2-DMH-induced SD rats, and 44 second generation rats of those induced rats was first preprocessed. Then, 12 feature variants were extracted using continuous wavelet analysis. The extracted feature variants were then inputted into the SVM for classification of normal, dysplasia, early carcinoma, and advanced carcinoma. Among the kernel functions the SVM used, the Poly and RBF kernels had the highest accuracy rates. The accuracy of the Poly kernel in normal, dysplasia, early carcinoma, and advanced carcinoma were 100, 97.5, 95% and 100% respectively. The accuracy of RBF kernel in normal, dysplasia, early carcinoma, and advanced carcinoma was 100, 95, 95% and 100% respectively. The results indicated that this method could effectively and easily diagnose colon cancer in its early stages.

A study on the early detection of colon cancer using the methods of wavelet feature extraction and SVM classifications of FTIR

This paper introduces a new method for the early detection of colon cancer using a combination of feature extraction based on wavelets for Fourier Transform Infrared Spectroscopy (FTIR) and classification using the Support Vector Machine (SVM). The FTIR data collected from 36 normal SD rats, 60 1,2-DMH-induced SD rats, and 44 second generation rats of those induced rats was first preprocessed. Then, 12 feature variants were extracted using continuous wavelet analysis. The extracted feature variants were then inputted into the SVM for classification of normal, dysplasia, early carcinoma, and advanced carcinoma. Among the kernel functions the SVM used, the Poly and RBF kernels had the highest accuracy rates. The accuracy of the Poly kernel in normal, dysplasia, early carcinoma, and advanced carcinoma were 100, 97.5, 95% and 100% respectively. The accuracy of RBF kernel in normal, dysplasia, early carcinoma, and advanced carcinoma was 100, 95, 95% and 100% respectively. The results indicated that this method could effectively and easily diagnose colon cancer in its early stages.