A, iPhone with dermatoscope attachment. B, Mobile app for image capture.
A and B, Images from patient 1 at initial and follow-up visits, respectively; no significant change was observed. C and D, Images from patient 2 at initial and follow-up visits, respectively; a significant change was observed.
Wu X, Oliveria SA, Yagerman S, Chen L, DeFazio J, Braun R, Marghoob AA. Feasibility and Efficacy of Patient-Initiated Mobile Teledermoscopy for Short-term Monitoring of Clinically Atypical Nevi. JAMA Dermatol. 2015;151(5):489-496. doi:10.1001/jamadermatol.2014.3837
Patient-driven mobile teledermoscopy may be applicable for monitoring of skin lesions.
To assess the feasibility, efficacy, and patient receptivity of teledermoscopy for short-term monitoring of clinically atypical nevi.
Design, Setting, and Participants
This was a prospective cohort study performed at an institutional referral center in New York. Consecutive patients 18 years or older, with 1 or more clinically atypical nevi that required short-term monitoring and were accessible by a mobile imaging device were recruited for the study. All 34 patients consented to the study, and 29 completed follow-up. Dermoscopic images were obtained in the office-based setting by a dermatologist and with an iPhone by the patient at baseline and follow-up (3-4 months). Patients completed surveys that included questions about skincare awareness and attitudes toward teledermoscopy. Standard dermoscopic images were evaluated by the office-based dermatologist, and mobile dermoscopic images were sent via the Internet to a teledermatologist to evaluate image quality and presence of significant clinical lesion change. The decisions of the teledermatologist and office-based dermatologist were compared.
Main Outcomes and Measures
(1) Feasibility of using mobile dermatoscope by patients, (2) diagnostic concordance of teledermoscopy vs conventional office-based visit, and (3) patient receptivity to teledermoscopy for short-term monitoring of nevi.
Of the 29 patients who completed the study, 28 (97%) were able to acquire baseline and follow-up images that were subsequently deemed evaluable by the teledermatologist. The diagnostic concordance between conventional office-based visits and teledermoscopy encounters was 0.87 (SE, 0.13) (κ statistic). In addition, patients reported high receptivity to teledermoscopy for short-term monitoring of nevi.
Conclusions and Relevance
Results from this pilot study suggest that teledermoscopy is feasible and effective as a method for short-term monitoring of clinically atypical nevi. The implementation of teledermoscopy can potentially enhance patient convenience, optimize physician scheduling, and promote efficiency.
Short-term dermoscopic monitoring is an established and accepted approach to monitor nonpalpable atypical melanocytic lesions.1,2 The standard interval for monitoring is 2.5 to 4.5 months, during which approximately 19% of dermoscopically monitored melanocytic neoplasms will manifest change.1 Of these changed lesions, 11% to 18% will eventually be diagnosed as malignant.1,2 Identifying dermoscopic change can be reliably performed by comparing dermoscopic images of the same lesion at 2 time points via side-by-side comparison.1,3 The presence of any clinically significant dermoscopic change, determined by a dermatologist, is deemed sufficient to warrant excision. If no change is present, the lesion is considered to be benign, obviating the need for a biopsy, and the patient can continue with clinical surveillance according to standard of care. The short-term monitoring approach to melanoma detection has high sensitivity1 and specificity1,2; most melanomas detected using this approach would be early stage, in situ, or low Breslow thickness (<1 mm).3
For more than 80% of patients who undergo short-term monitoring,1 no changes of the atypical lesion will be detected, and no further clinical interventions will be required. The simplicity of the follow-up visits for most patients makes short-term monitoring ideally suited for store-and-forward teledermoscopy, whereby patients capture and transmit an electronic image of the monitored lesion via a mobile dermatoscope to the dermatologist for evaluation. For this process to be practical and acceptable, it is necessary to show that patient-acquired dermoscopic images are of sufficient quality and that remote follow-up evaluation has high concordance with a clinical evaluation. We conducted a pilot study to assess the potential of using patient-initiated mobile teledermoscopy for short-term monitoring of clinically atypical pigmented lesions. We evaluated patient-initiated mobile teledermoscopy with respect to feasibility, efficacy (as measured by diagnostic concordance with conventional office-based visits) and receptivity.
Participants were recruited from the practices of 2 dermatologists (J.D. and A.A.M.) in the Dermatology Service at Memorial Sloan Kettering Cancer Center (MSKCC) from February 2013 through February 2014.
Eligible participants were 18 years or older with 1 or more atypical pigmented lesions deemed suitable for short-term dermoscopic monitoring by the dermatologist. Patients were excluded if they could not fully understand or participate in the informed consent process, had physical limitations hindering the use of a mobile device, had lesions on body locations that were difficult to self-monitor, and did not have a family member willing to take the image. Participation was voluntary, and no compensation was provided. All patients provided written informed consent. Approval for this study was obtained from the institutional review board at MSKCC.
The dermatologist identified the lesion for monitoring and then instructed the patient regarding the lesion of interest. Baseline and follow-up visits were conducted with an interval of 3 to 4 months.1,2 At both visits, 2 sets of images were obtained for the lesion of interest: 1 “standard-of-routine” dermoscopic image taken by a clinic professional using a Nikon D90 single-lens reflex digital camera and 60-mm Macro Nikkor lens (Nikon) with a dermoscopy attachment (EpiLume; Canfield Imaging Systems Inc) and 1 by the patient with a commercially available “mobile” dermatoscope (DermScope; Canfield Imaging Systems Inc; http://www.canfieldsci.com/imaging_systems/dermatoscopes/VEOS_DS3.html) attached to a standard mobile phone equipped with a built-in 8-megapixel camera (iPhone 4S; Apple Inc). The mobile dermatoscope had an installed application that facilitated the image capture and transmission process (DermScope; Canfield Imaging Systems Inc; downloaded from iTunes, Apple Inc) (Figure 1). The mobile phones and dermatoscopes were the property of MSKCC and were used by the patients prior to their office visits. Both the standard-of-routine and mobile dermoscopic images were obtained under contact, polarized light with original magnification ×10 magnification. The patients were given the same printed instruction sheet at both initial and follow-up visits to guide them through the process of coupling the dermatoscope to the phone and acquiring the image. At the initial visit, patients received verbal guidance about the instruction sheet and how to use the device. At the follow-up visit, patients received no verbal guidance to better simulate the home environment, where they would be expected to conduct the teledermoscopy process independently.
Patients were asked to complete surveys at baseline and follow-up visits, after the acquisition of images. The surveys included questions about their attitudes toward mobile teledermoscopy based on their encounter. The baseline survey also included general questions about skincare awareness, and the follow-up survey included questions about factors that would influence patient use of teledermoscopy and the amount that patients were willing to pay out of pocket for a mobile dermatoscope. Family members enrolled in the study also completed the survey.
During the follow-up visit, the office-based dermatologist (A.A.M.) evaluated the monitored lesion based on clinical information and side-by-side standard-of-routine dermoscopic images at 3-million megapixel resolution stored in a secure database (DermaGraphix; Canfield Imaging Systems Inc). Significant change in a monitored lesion was defined as any visual changes except global darkening or lightening consistent with surrounding skin tanning changes and increase or decrease in the number of milia-like cysts1,2 Clinical management decisions were based on the standard-of-routine visits. The saved mobile images from the initial and the follow-up visits were placed side by side and e-mailed via a secure network in jpeg format to an outside expert dermatologist (R.B.) for teleconsultation. Figure 2 illustrates examples of side-by-side dermoscopic images captured by the patients. The same evaluation criteria for significant change used by the office-based dermatologist were used by the teledermatologist, who also assessed the quality of the images, whether they were in focus, and were of sufficient quality to allow for evaluation. No clinical data regarding the patients were sent to the teledermatologist.
To obtain information on feasibility, the research study assistant completed an evaluator checklist for each patient based on observation of the patient, to characterize the level of difficulty encountered during image acquisition. The checklist consisted of 4 items that assessed individual stages of the image acquisition process, from assembling the mobile dermatoscope to capturing the image via the mobile app.
Efficacy was determined by measuring diagnostic concordance between the clinical decisions of the teledermatologist and the office-based dermatologist. The diagnostic end point was defined as significant change of lesion warranting biopsy during the monitoring interval.
Patient receptivity to teledermoscopy was assessed at initial and follow-up visits. A survey was administered to the patients and included questions about patient confidence in their ability to use the mobile dermatoscope and barriers and facilitators to the use of the mobile dermatoscope.
Descriptive statistics were used to characterize the feasibility measures. The proportion of study participants who successfully completed all steps of the image acquisition process and the proportion of participants who captured at least 1 pair of useable (in-frame and in-focus) images were determined.
For patient receptivity, responses to each survey question were coded on a Likert scale of 1 to 5 (1, not at all; 2, a little; 3, somewhat; 4, very; 5, extremely) and descriptive statistics were used to calculate overall means (SDs) and proportions for each response, at baseline and follow-up with the domain category. Paired t test was used to analyze results between initial and follow-up visits.
To assess the interrater agreement between conventional office-based diagnosis and diagnosis made via teledermoscopy, percentage of agreement and κ statistics were calculated. The κ statistic was interpreted using the following cutoff points as agreed on in the literature: less than 0.2, poor agreement; 0.21 to 0.4, fair; 0.41 to 0.6, moderate; 0.61 to 0.8, strong; and more than 0.8, near-complete agreement.4
A total of 34 patients were recruited and consented to participate (18 women and 16 men; mean age, 43.6 years [range, 18-81 years]); 29 completed follow-up. Of the remaining 5 patients, 3 failed to complete the study, 1 failed to save the mobile image at follow-up, and 1 was ineligible because the family member who participated during the initial visit was not present at follow-up. A total of 33 lesions were eligible for monitoring among the 29 patients (4 patients had 2 nevi monitored).
Participants had an overall basic knowledge about skin cancer, with 91% knowing that there is more than 1 type of skin cancer and 100% knowing that skin cancer is curable when diagnosed early. Eighty-eight percent of the participants had experience with a mole that was enough of a cause for concern to their dermatologist to require physician monitoring. Patients also demonstrated a high level of self-awareness of their skin: 91% reported they pay attention to their skin, and 69% reported they would notice a changing lesion. While 81% of patients endorsed bringing a concerning lesion to a physician’s attention, only 9% reported ever having consulted a physician via teledermatology.
All 33 mobile dermoscopic image pairs (100%) taken on initial and follow-up visits were within frame, and 29 (88%) were in-focus. Of the 4 image pairs with an image out of focus, only 2 pairs were deemed to be of insufficient quality for evaluation. Of the 29 patients who had completed follow-up, 28 (97%) had at least 1 image pair that was both in frame and in focus (Table 1).
Results from clinical decisions made for the 30 image pairs evaluated by the teledermatologist showed that 97% were in agreement with the decision made by the office-based dermatologist (25 had no significant change, and 4 had significant change). The 1 discordant case was diagnosed by the teledermatologist as change, whereas the office-based dermatologist made a diagnosis of no change. The κ statistic was calculated to be 0.87 (SE, 0.13), which was deemed to be near-complete agreement.
The 1 discordant case could potentially be attributed to a lower threshold for change by the teledermatologist owing to diminished image quality to prevent overlooking minute or ambiguous changes in monitored lesions.
Results for patient receptivity for confidence in the ability to use mobile dermatoscope for image capture (domain 1), facilitators for use (domain 2), and barriers to use (domain 3) are presented in Tables 2, 3, and 4.
Questions in domains 1 and 2 reflect patient user confidence and favorable attitude toward teledermoscopy; patient agreement in these domains was significantly higher compared with responses for questions in domain 3, which represent a negative attitude toward teledermoscopy (P < .01). There were no statistically significant differences in responses between the initial and follow-up visits. Overall, patients reported a high degree of confidence in their ability to use the mobile dermatoscope to capture images (means [SDs], 4.57 [0.64] and 4.62 [0.49]), on a scale of 1 to 5, at initial and follow-up visits, respectively) and strong favorability to use the mobile dermatoscope (4.59 [0.86] and 4.46 [0.84]). Most patients did not report barriers to use of the dermatoscope (1.85 [1.23] and 1.82 [1.26] on a scale of 1-5). Of the listed reasons against the use of a mobile dermatoscope, the patients’ desire to see a dermatologist in office ranked highest compared with other barriers to use (2.85 [1.35] and 3.04 [1.31]). Technical aspects of using the mobile dermatoscope were not a concern to most patients.
The most frequently cited reasons by patients for use of teledermoscopy to monitor their lesions were shorter wait time for appointment, potential for more frequent monitoring, and improved privacy and comfort level. A few patients mentioned cost and concern about insurance reimbursement as a potential barrier to use. Patients indicated that they would be willing to pay out of pocket to acquire the mobile dermatoscope and spend from $20 to $500 (median amount, $100).
Teledermatology has been successfully used for remote diagnosis and consultation.5- 7 Teledermoscopy is a specialized approach within teledermatology that is more reliant on image quality as well as the dermoscopic expertise of the evaluating dermatologists and has the potential for improved diagnostic accuracy compared with teledermatology without dermoscopy.8,9 The potential use of mobile teledermoscopy in skin cancer screening was first introduced by Massone et al10 in 2007, opening new possibilities in the area of teledermatology. The present work is, to our knowledge, the first study to evaluate the feasibility, efficacy, and receptivity of teledermoscopy for the short-term monitoring of pigmented lesions. Our results showed that the use of teledermoscopy in short-term monitoring is highly feasible, has strong diagnostic concordance with conventional clinical visits, and is well received by patients.
Prior studies6,8- 13 that assessed the diagnostic reliability of store-and-forward teledermoscopy for skin cancer triage and diagnosis of pigmented skin tumors have reported high concordance rates, similar to those observed in our study. Reported diagnostic concordances of teledermoscopy to standard clinical visits have ranged from 0.63 to 0.91 (κ statistic).13- 15
Patient satisfaction with store-and-forward teledermatology has been reported.5,16- 21 Lim et al21 conducted a survey study to evaluate patient receptivity of teledermoscopy in 200 patients after they received care via a commercial virtual lesion clinic in New Zealand. Eighty two percent of the surveyed patients reported confidence in the teledermoscopy service, and 73% preferred their teledermoscopy encounter to conventional office visits. In another study, Whited et al20 conducted a randomized clinical trial in patients needing dermatology consultations at Durham Veterans Affairs Medical Center, in which 135 patients were randomized to teledermatology and 140 patients to conventional office consultations. Surveys were used to evaluate patient experience. Eighty four percent of patients randomized to teledermatology reported confidence in their teleconsultation experience, and 66% thought it was more convenient than conventional office consultations. Other studies have reported similarly high patient acceptance rates: Kvedar et al18 found the average patient satisfaction to be 4.56 out of 5, and van den Akker et al19 reported a rate of 7.4 out of 10 in patients who underwent teledermatology consultations referred by their primary care physicians. Our results are consistent with those of prior studies and are unique in that they are the first to report patient receptivity of teledermoscopy, where the patients were responsible for the image acquisition process. Patient-initiated teledermoscopy involves an arguably more intricate image acquisition process than in regular teledermatology owing to the higher emphasis on image quality. To our knowledge, only 1 prior pilot study by Janda et al22 has investigated patient-initiated teledermoscopy. They tested the feasibility of using patient-initiated teledermoscopy during self–skin-examination in 10 volunteers. Similar to our results, they found that the participants were able to learn how to use the technology without much difficulty.22 Our patients reported high self-confidence in their ability to use the mobile teledermatoscope and strongly agreed with the reasons that support the use of teledermoscopy. In particular, the patients thought the mobile dermatoscope was easy to use, reduced the wait time of appointments, saved them a trip to the physician’s office, and provided added comfort and privacy. Among the concerns that patients had was reliability of diagnosis from teledermatologist; part of this anxiety could potentially be alleviated if the patients had an established relationship with or had direct access to the remote teledermatologist.
Limitations of this study include the small sample size and the fact that the image capture via mobile dermatoscope was performed in the office with a research assistant present, which may have affected patients’ confidence level in their ability to use the mobile dermatoscope more than if it were performed independently by patients at home. Also, mobile images were saved and transmitted for teleconsultation by the research assistant, which eliminates some of the obstacles that patients might have encountered when they performed the process independently. However, image transmission is facilitated with the mobile app specifically designed for this purpose. The captured image can be saved and transmitted by the click of a button on the touch screen. In addition, our study population comprised high-risk patients who are more motivated in skin care than the general population and are potentially more receptive to incorporating new technologies into their care, limiting the generalizability of the results. Another potential limitation in analyzing sequential images revolves around image orientation, color calibration, and exposure. The software used in this study afforded us the ability to rotate images so that the orientation matches one another, and thus mismatched orientation of images proved not to be a limitation in our study. However, color calibration is a problem that plagues virtually all systems and cameras. This dilemma exists irrespective of what system is used. We do not have a solution to this problem. While the color calibration issue is an important one that we and the industry need to tackle, the lack of precise color calibration has never presented a problem for us when it comes to short-term mole monitoring. The bigger problem is differences in exposure (overall lighter or darker images), but fortunately this is easily fixed by matching the exposure for 2 images, a process that the software we used handled quite easily.
In this small-scale pilot study, we have demonstrated that patients are both capable of and receptive to the use of mobile teledermoscopy in short-term monitoring of their pigmented lesions. Teledermatology expands the portals of communication between patient and physician; in addition to facilitating remote diagnosis, it could open doors for patient-driven home monitoring systems. Some dermatologic conditions, such as psoriasis17 and chronic leg ulcers,23 have been successfully monitored by patients and their physicians via store-and-forward mobile teledermatology. Results of a recent pilot study24 suggested that patients with multiple nevi who are at high risk for melanoma may benefit from mobile teledermoscopy in monitoring of atypical lesions preselected by dermatologists, thereby decreasing the number of office visits and optimizing physician care.
Facing an already-large patient population and an anticipated increase in new patients as a result of the Affordable Care Act, the demand for access to dermatologic care and a shortage of resources is becoming considerable.25 The estimated wait time for a dermatology appointment for a new patient already exceeds 30 days and is only expected to worsen. Store-and-forward teledermatologic care could provide a cost-effective and time-efficient solution to the problem.25
Some of the concerns raised around teledermatology include patient privacy, the physician-patient relationship, and patient ability to select lesions of medical concern lesions for imaging. Our ultimate aim is to propose a delivery care system whereby short-term monitoring incorporates teledermoscopy and also addresses these potential downsides. Under our modality of care, patients needing short-term monitoring will have an established relationship with their dermatologists, who will be the ones identifying concerning lesions that need to be monitored and the ones who evaluate the lesions via teledermoscopy and communicate treatment options directly with the patients. This system has the benefit of maintaining the patient-physician relationship while optimizing efficiency of care. If successfully implemented, the system is expected to enhance patient compliance, confidence, and efficacy in skin self-photography while optimizing physician access for the subset of lesions that is of the most cause for concern. In turn, physicians will benefit from improved patient satisfaction, optimized patient scheduling, and enrichment of procedural office visits leading to improved skin cancer detections. Furthermore, this system has the added benefit of reducing cost and augmenting economic efficiency at a time of increasing scarcity of health care resources. The findings from this study provide important preliminary data that will allow the advancement of research in an evolving new front of medicine: patient-driven care.
Corresponding Author: Ashfaq A. Marghoob, MD, Dermatology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, Hauppauge, 160 E 53rd St, Second Floor, New York, NY 10022 (email@example.com).
Accepted for Publication: September 14, 2014.
Published Online: January 28, 2015. doi:10.1001/jamadermatol.2014.3837.
Author Contributions: Ms Wu, Dr Oliveria, and Dr Marghoob had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Oliveria, Chen, Braun, Marghoob.
Acquisition, analysis, or interpretation of data: Wu, Oliveria, Yagerman, DeFazio, Braun, Marghoob.
Drafting of the manuscript: Wu, Oliveria.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Wu, Oliveria.
Administrative, technical, or material support: Wu, Yagerman, Chen, DeFazio, Marghoob.
Study supervision: Oliveria, Marghoob.
Conflict of Interest Disclosures: None reported.