Pre-diagnostic digital imaging prediction model to discriminate between malignant melanoma and benign pigmented skin lesion

Automatic Skin Cancer Detection Using Clinical Images: A Comprehensive Review

Skin cancer has become increasingly common over the past decade, with melanoma being the most aggressive type. Hence, early detection of skin cancer and melanoma is essential in dermatology. Computational methods can be a valuable tool for assisting dermatologists in identifying skin cancer. Most research in machine learning for skin cancer detection has focused on dermoscopy images due to the existence of larger image datasets. However, general practitioners typically do not have access to a dermoscope and must rely on naked-eye examinations or standard clinical images. By using standard, off-the-shelf cameras to detect high-risk moles, machine learning has also proven to be an effective tool. The objective of this paper is to provide a comprehensive review of image-processing techniques for skin cancer detection using clinical images. In this study, we evaluate 51 state-of-the-art articles that have used machine learning methods to detect skin cancer over the past decade, focusing on clinical datasets. Even though several studies have been conducted in this field, there are still few publicly available clinical datasets with sufficient data that can be used as a benchmark, especially when compared to the existing dermoscopy databases. In addition, we observed that the available artifact removal approaches are not quite adequate in some cases and may also have a negative impact on the models. Moreover, the majority of the reviewed articles are working with single-lesion images and do not consider typical mole patterns and temporal changes in the lesions of each patient.

Computer-assisted diagnosis techniques (dermoscopy and spectroscopy-based) for diagnosing skin cancer in adults

BACKGROUND

Early accurate detection of all skin cancer types is essential to guide appropriate management and to improve morbidity and survival. Melanoma and cutaneous squamous cell carcinoma (cSCC) are high-risk skin cancers which have the potential to metastasise and ultimately lead to death, whereas basal cell carcinoma (BCC) is usually localised with potential to infiltrate and damage surrounding tissue. Anxiety around missing early curable cases needs to be balanced against inappropriate referral and unnecessary excision of benign lesions. Computer-assisted diagnosis (CAD) systems use artificial intelligence to analyse lesion data and arrive at a diagnosis of skin cancer. When used in unreferred settings ('primary care'), CAD may assist general practitioners (GPs) or other clinicians to more appropriately triage high-risk lesions to secondary care. Used alongside clinical and dermoscopic suspicion of malignancy, CAD may reduce unnecessary excisions without missing melanoma cases.

OBJECTIVES

To determine the accuracy of CAD systems for diagnosing cutaneous invasive melanoma and atypical intraepidermal melanocytic variants, BCC or cSCC in adults, and to compare its accuracy with that of dermoscopy.

SEARCH METHODS

We undertook a comprehensive search of the following databases from inception up to August 2016: Cochrane Central Register of Controlled Trials (CENTRAL); MEDLINE; Embase; CINAHL; CPCI; Zetoc; Science Citation Index; US National Institutes of Health Ongoing Trials Register; NIHR Clinical Research Network Portfolio Database; and the World Health Organization International Clinical Trials Registry Platform. We studied reference lists and published systematic review articles.

SELECTION CRITERIA

Studies of any design that evaluated CAD alone, or in comparison with dermoscopy, in adults with lesions suspicious for melanoma or BCC or cSCC, and compared with a reference standard of either histological confirmation or clinical follow-up.

DATA COLLECTION AND ANALYSIS

Two review authors independently extracted all data using a standardised data extraction and quality assessment form (based on QUADAS-2). We contacted authors of included studies where information related to the target condition or diagnostic threshold were missing. We estimated summary sensitivities and specificities separately by type of CAD system, using the bivariate hierarchical model. We compared CAD with dermoscopy using (a) all available CAD data (indirect comparisons), and (b) studies providing paired data for both tests (direct comparisons). We tested the contribution of human decision-making to the accuracy of CAD diagnoses in a sensitivity analysis by removing studies that gave CAD results to clinicians to guide diagnostic decision-making.

MAIN RESULTS

We included 42 studies, 24 evaluating digital dermoscopy-based CAD systems (Derm-CAD) in 23 study cohorts with 9602 lesions (1220 melanomas, at least 83 BCCs, 9 cSCCs), providing 32 datasets for Derm-CAD and seven for dermoscopy. Eighteen studies evaluated spectroscopy-based CAD (Spectro-CAD) in 16 study cohorts with 6336 lesions (934 melanomas, 163 BCC, 49 cSCCs), providing 32 datasets for Spectro-CAD and six for dermoscopy. These consisted of 15 studies using multispectral imaging (MSI), two studies using electrical impedance spectroscopy (EIS) and one study using diffuse-reflectance spectroscopy. Studies were incompletely reported and at unclear to high risk of bias across all domains. Included studies inadequately address the review question, due to an abundance of low-quality studies, poor reporting, and recruitment of highly selected groups of participants.Across all CAD systems, we found considerable variation in the hardware and software technologies used, the types of classification algorithm employed, methods used to train the algorithms, and which lesion morphological features were extracted and analysed across all CAD systems, and even between studies evaluating CAD systems. Meta-analysis found CAD systems had high sensitivity for correct identification of cutaneous invasive melanoma and atypical intraepidermal melanocytic variants in highly selected populations, but with low and very variable specificity, particularly for Spectro-CAD systems. Pooled data from 22 studies estimated the sensitivity of Derm-CAD for the detection of melanoma as 90.1% (95% confidence interval (CI) 84.0% to 94.0%) and specificity as 74.3% (95% CI 63.6% to 82.7%). Pooled data from eight studies estimated the sensitivity of multispectral imaging CAD (MSI-CAD) as 92.9% (95% CI 83.7% to 97.1%) and specificity as 43.6% (95% CI 24.8% to 64.5%). When applied to a hypothetical population of 1000 lesions at the mean observed melanoma prevalence of 20%, Derm-CAD would miss 20 melanomas and would lead to 206 false-positive results for melanoma. MSI-CAD would miss 14 melanomas and would lead to 451 false diagnoses for melanoma. Preliminary findings suggest CAD systems are at least as sensitive as assessment of dermoscopic images for the diagnosis of invasive melanoma and atypical intraepidermal melanocytic variants. We are unable to make summary statements about the use of CAD in unreferred populations, or its accuracy in detecting keratinocyte cancers, or its use in any setting as a diagnostic aid, because of the paucity of studies.

AUTHORS' CONCLUSIONS

In highly selected patient populations all CAD types demonstrate high sensitivity, and could prove useful as a back-up for specialist diagnosis to assist in minimising the risk of missing melanomas. However, the evidence base is currently too poor to understand whether CAD system outputs translate to different clinical decision-making in practice. Insufficient data are available on the use of CAD in community settings, or for the detection of keratinocyte cancers. The evidence base for individual systems is too limited to draw conclusions on which might be preferred for practice. Prospective comparative studies are required that evaluate the use of already evaluated CAD systems as diagnostic aids, by comparison to face-to-face dermoscopy, and in participant populations that are representative of those in which the test would be used in practice.

Deep-learning-based, computer-aided classifier developed with a small dataset of clinical images surpasses board-certified dermatologists in skin tumour diagnosis

BACKGROUND

Application of deep-learning technology to skin cancer classification can potentially improve the sensitivity and specificity of skin cancer screening, but the number of training images required for such a system is thought to be extremely large.

OBJECTIVES

To determine whether deep-learning technology could be used to develop an efficient skin cancer classification system with a relatively small dataset of clinical images.

METHODS

A deep convolutional neural network (DCNN) was trained using a dataset of 4867 clinical images obtained from 1842 patients diagnosed with skin tumours at the University of Tsukuba Hospital from 2003 to 2016. The images consisted of 14 diagnoses, including both malignant and benign conditions. Its performance was tested against 13 board-certified dermatologists and nine dermatology trainees.

RESULTS

The overall classification accuracy of the trained DCNN was 76·5%. The DCNN achieved 96·3% sensitivity (correctly classified malignant as malignant) and 89·5% specificity (correctly classified benign as benign). Although the accuracy of malignant or benign classification by the board-certified dermatologists was statistically higher than that of the dermatology trainees (85·3% ± 3·7% and 74·4% ± 6·8%, P < 0·01), the DCNN achieved even greater accuracy, as high as 92·4% ± 2·1% (P < 0·001).

CONCLUSIONS

We have developed an efficient skin tumour classifier using a DCNN trained on a relatively small dataset. The DCNN classified images of skin tumours more accurately than board-certified dermatologists. Collectively, the current system may have capabilities for screening purposes in general medical practice, particularly because it requires only a single clinical image for classification.

Performance of residents using digital images versus glass slides on certification examination in anatomical pathology: a mixed methods pilot study

BACKGROUND

It is anticipated that many licensing examination centres for pathology will begin fully digitizing the certification examinations. The objective of our study was to test the feasibility of a fully digital examination and to assess the needs, concerns and expectations of pathology residents in moving from a glass slide-based examination to a fully digital examination.

METHODS

We conducted a mixed methods study that compared, after randomization, the performance of senior residents (postgraduate years 4 and 5) in 7 accredited anatomical pathology training programs across Canada on a pathology examination using either glass slides or digital whole-slide scanned images of the slides. The pilot examination was followed by a post-test survey. In addition, pathology residents from all levels of training were invited to participate in an online survey.

RESULTS

A total of 100 residents participated in the pilot examination; 49 were given glass slides instead of digital images. We found no significant difference in examination results between the 2 groups of residents (estimated marginal mean 8.23/12, 95% confidence interval [CI] 7.72-8.87, for glass slides; 7.84/12, 95% CI 7.28-8.41, for digital slides). In the post-test survey, most of the respondents expressed concerns with the digital examination, including slowly functioning software, blurring and poor detail of images, particularly nuclear features. All of the respondents of the general survey (n = 179) agreed that additional training was required if the examination were to become fully digital.

INTERPRETATION

Although the performance of residents completing pathology examinations with glass slides was comparable to that of residents using digital images, our study showed that residents were not comfortable with the digital technology, especially given their current level of exposure to it. Additional training may be needed before implementing a fully digital examination, with consideration for a gradual transition.

Recurrent papillary urothelial neoplasm of low malignant potential. Subtle architectural disorder detected by quantitative analysis in DAXX-immunostained tissue sections

The aim of the study was to identify subtle changes in the so-called architectural predominant order in nonrecurrent and recurrent papillary urothelial neoplasm of low malignant potential (PUNLMP). Quantitative analysis was performed with a software package written in LabVIEW (National Instruments, Austin, TX) in DAXX-immunostained tissue sections. Twelve cases of PUNLMP with papillary fronds sectioned lengthwise through the core were investigated and subdivided as follows: 7 nonrecurrent and 5 recurrent PUNLMP cases. Six cases of normal urothelium (NU) were included. When an epithelial thickness threshold is set at 108 μm (ie, 400 pixels), there is a complete separation between NU and PUNLMP; however, nonrecurrent and recurrent cases fall in the same range of thickness. In setting a nuclear elongation factor threshold at 2.1, there are differences between the 2 PUNLMP groups, recurrent PUNLMP and NU cases, showing a somewhat similar proportion of elongated nuclei. The nuclear orientation separates nonrecurrent from recurrent PUNLMP groups; however, NU cases do not appear as a separate group from the 2 PUNLMP groups. In combining epithelial thickness, nuclear elongation, and orientation in a multivariate analysis, the 2 PUNLMP groups appear separate between them and from NU. NU is less thickened than the 2 PUNLMP groups and shows a combination of elongated and less elongated nuclei. Elongated nuclei are more numerous in nonrecurrent PUNLMP, whereas the nuclei in recurrent PUNLMP are less elongated and less polarized than in the other group. Such finding can be used routinely to identify those PUNLMP patients who will have recurrence.

Computer-aided diagnosis of skin lesions using conventional digital photography: a reliability and feasibility study

BACKGROUND

Computer-aided diagnosis (CADx) software that provides a second opinion has been widely used to assist physicians with various tasks. In dermatology, however, CADx has been mostly limited to melanoma or melanocytic skin cancer diagnosis. The frequency of non-melanocytic skin cancers and the accessibility of regular digital macrographs have raised interest in developing CADx for broader applications.

OBJECTIVES

To investigate the feasibility of using CADx to diagnose both melanocytic and non-melanocytic skin lesions based on conventional digital photographic images.

METHODS

This study was approved by an institutional review board, and the requirement to obtain informed consent was waived. In total, 769 conventional photographs of melanocytic and non-melanocytic skin lesions were retrospectively reviewed and used to develop a CADx system. Conventional and new color-related image features were developed to classify the lesions as benign or malignant using support vector machines (SVMs). The performance of CADx was compared with that of dermatologists.

RESULTS

The clinicians' overall sensitivity, specificity, and accuracy were 83.33%, 85.88%, and 85.31%, respectively. New color correlation and principal component analysis (PCA) features improved the classification ability of the baseline CADx (p = 0.001). The estimated area under the receiver operating characteristic (ROC) curve (Az) of the proposed CADx system was 0.949, with a sensitivity and specificity of 85.63% and 87.65%, respectively, and a maximum accuracy of 90.64%.

CONCLUSIONS

We have developed an effective CADx system to classify both melanocytic and non-melanocytic skin lesions using conventional digital macrographs. The system's performance was similar to that of dermatologists at our institute. Through improved feature extraction and SVM analysis, we found that conventional digital macrographs were feasible for providing useful information for CADx applications. The new color-related features significantly improved CADx applications for skin cancer.

Near-infrared tissue oximetry and digital image analysis: quantification of renal ischaemia in real time during partial nephrectomy

OBJECTIVE

To determine the feasibility of using near-infrared tissue oximetry (TO) and digital image analysis for assessing renal function and to quantify local renal ischaemia in a porcine model.

MATERIALS AND METHODS

Tissue oximetry was performed and red/blue (R/B) colour ratios were determined on renal units of Yorkshire swine. Interval measurements were taken before clamping the renal hilum, during warm ischaemia, and after unclamping using a ViOptix T.Ox™ Tissue Oximeter and Matlab(®) digital image analysis. Matlab software analysed images from the laparoscopic camera and determined an R/B ratio to track renal ischaemia. The tissue oximeter used direct infrared light and was placed adjacent to the kidney parenchymal surface. The data were divided into preclamp, clamp and post-clamp, and compared between methodologies.

RESULTS

The R/B ratio showed a higher rate of change compared with TO during clamp time in both the 15-min experiment (R/B = 96.0 vs. TO = 52.1 unit/reference) and the 30-min experiment (R/B = 97.6 vs. TO = 45.9). The R/B ratio showed a higher rate of change compared with TO at 1 min after clamping in the 15-min experiment (R/B = 80.1 vs. TO = 12.4). Both detection devices showed similar changes in pre- and post-clamp measurements in the 15- min experiment (R/B = 1.6 vs. TO = 3.8) and the 30-min experiment (R/B = 4.7 vs. TO =-4.5). In the 30-min experiment the R/B ratio showed a significant difference between preclamp, clamp, and post-clamp states (P= 0.026).

CONCLUSIONS

Both TO and digital image analysis were able to calculate an ischaemic drop in tissue oxygen saturation during periods of acute renal ischaemia. The findings suggest that the R/B ratio observed during histogram analysis shows a greater sensitivity compared with TO in quantifying renal ischaemia.