Reliability of Digital Image-Based Psoriasis Area and Severity Index Scores

In this interview, David Wu, PhD, met with The Dermatologist to discuss the reliability of digital image-based Psoriasis Area and Severity Index (DIB-PASI) scores for patients of different skin tones analyzed by experienced and trained clinical accessors. Dr Wu completed his doctoral work in the laboratory of Daniel Lim, MD, PhD, where he leveraged large-scale genomic screens and single-cell RNA-sequencing technologies to interrogate human stem cell differentiation. During his clinical years at the University of California San Francisco (UCSF), Dr Wu has worked with Wilson Liao, MD, an expert in the genetics of psoriasis and professor and vice chairperson of research in the department of dermatology. Their current work spans research in teledermatology, as well as massive-scale genomic studies of patients with psoriasis undergoing biologic treatment.

The Dermatologist: Can you elaborate on the reliability of face-to-face versus DIB-PASI scores?
Dr Wu: Because psoriasis is a chronic disease, we need to reliably track it in our patients over time. To do this systematically, the PASI score has become a standard way of communicating the extent of a patient’s disease. Typically, a dermatologist would calculate this in person when they see the patient but with the advent of telehealth and widespread digital photography, many dermatologists wondered, can we do the same scoring from photographs of the patient and get accurate scores?
     We were inspired by pioneering studies in Europe and the United States that looked at this question to conduct our own pilot study in our own patient population, but we added a few elements. One element was different skin tones. We checked whether the DIB-PASI score was equally reliable depending on the patient’s skin tone. And then a second element was to see whether the experience level of the dermatologists who were doing these remote assessments also impacted reliability. We found that the DIB-PASI score was still reliable, no matter the patient’s skin tone or the training level of the assessor.

The Dermatologist: Part of your investigation was to evaluate the reliability of DIB-PASI scores in patients with skin of color. Can you tell us a bit about your results in this area?
Dr Wu: Most dermatologists are trained on photos of lighter skin, so they are experienced at evaluating patients with light skin who have the classic presentation of psoriasis. This is more challenging for patients with darker skin tones, where psoriasis can look diff erent even if it is equally severe. In our study, we broke patients down into 2 simple categories: lighter skin tones and medium to dark skin tones. We found that the reliability of the DIB-PASI score was comparable for both groups. There was no compromise or impact, and the DIB-PASI score was still a robust way of assessing our patients regardless of their skin tone.
    I do want to note that there can be in-person biases in evaluating psoriasis in patients with skin of color, and it is possible that these same biases translate to evaluating photographs. You might get the reliability and similar scores, but there may still be an inherent bias that we need to address as a community. Our study is not a panacea for that, but at least we know we are not systematically adding new biases that would compromise scoring for any of our patients.

The Dermatologist: In your study, you note that the findings lay the foundation for expanding teledermatology-based clinical trials for patients with psoriasis. Can you elaborate on what ways this can be useful to dermatologists?
Dr Wu: In clinical trials of psoriasis, especially of new therapies, the PASI score has become a go-to metric that is often used and reported. If you want to do a clinical trial, you are going to be assessing your patients frequently and getting PASI scores to see how they are responding to treatment. For patients who do not live near an urban academic medical center, getting these frequent in-person assessments is not feasible. In fact, a lot of patients do not participate in clinical trials because it is not feasible for them to come for so many assessments in person.
     For example, even at UCSF we have patients who drive 3 or 4 hours each way just to come to our psoriasis center for a single appointment. And if they were to do that repeatedly, they are losing days of work and that is not something they can do. If we can show that we could do this remotely in a reliable way and convert even some of these appointments into remote assessments, it could transform how this entire population is participating in clinical trials. Patients in remote areas have been traditionally ignored and underserved simply because of that lack of feasibility. We think the DIB-PASI score could supercharge clinical trials if it is truly reliable.

The Dermatologist: What further research is needed to assess the reliability of DIB-PASI scores across patients with skin of color?
Dr Wu: There are 3 key parts that we are thinking about. First, larger studies of diverse patient cohorts are needed. Our pilot study was small. It was promising, but we want to expand this to see a lot of patients and good representation in different backgrounds. And this will even empower us to look at other factors that might be impacting reliability that we do not even know about, such as age, sex, gender, and weight. All these things can be distorted in a photo. We hope that these factors are not limiting reliability, but we need to answer those questions.
     The second part is our dream of patients being able to use their own smartphones at home to easily capture standardized photographs. The challenge, of course, is that you do not necessarily get standardized or clear photographs from patients all the time. We do think that with proper guidance on how to take a good photo, as well as the right software, we can help make this a reality.
     The final part is the big structural challenge that needs to be resolved to make headway in this area. As I mentioned before, most of us are trained to see psoriasis in lighter skin tones and so we have less experience when we see it in patients with darker skin tones and might not be able to evaluate it as reliably. This is something as a community we have been working on, and there is still a long way to go. We hope that if we can get patients to take standardized photographs in their own homes that are high enough quality to use in a clinical trial, they will be high enough quality to use for education as well. We can deidentify the photos of patients who are willing to donate them and create better sets of images of psoriasis in the real world.

The Dermatologist: Do you foresee challenges regarding DIB-PASI scores and privacy issues?
Dr Wu: Privacy, security, and protecting patient health information are so critical for health care as a whole and particularly dermatology, which has so many photos, especially photos of sensitive areas that patients would not want to get out on the internet. As we develop these tools, it is not acceptable for privacy and security to be breached. We need to make sure we protect the privacy of our patients. With photographs, we can accurately deidentify them and if it is a close-up image of a rash or a part of the skin, it can be hard to tell who or where it is from. This is a built-in kind of privacy, but we cannot rely on that alone. People will have identifiable tattoos, or jewelry, or things that could link them to their health record. We need to make sure that this is completely safe, secure, and private to protect our patients.

The Dermatologist: Do you foresee any other challenges for the use of DIB-PASI scores?
Dr Wu: There is a general hesitancy in the fi eld of medicine to adopt new technologies, and that is for a great reason. We do not want to just start doing something new that impacts our patients without knowing it is safe and reliable. Many dermatologists and patients may believe that an in-person visit is the gold standard and a photograph is never going to be as good as being seen face to face. I can totally understand that feeling. It is up to the community to do this research to find out how reliable the DIB-PASI score is. Studies are showing some promising results, but we really need to expand these to larger studies and look at all the factors to ensure that this is reliable for everyone.

The Dermatologist: Are there any other insights you would like to share with your colleagues regarding the reliability of DIB-PASI scores?
Dr Wu: If we can use photographs to reliably evaluate skin disease in all different skin tones, we could transform clinical trials, not just in psoriasis, but perhaps for many other dermatologic conditions. Including patients who have been traditionally underrepresented and underserved in our clinical research is critical for advancing health care for all. The core technologies exist and it is a matter of putting together the pieces of this complicated puzzle to make it work well. We are excited about the possibility that we can bring high-quality health care and clinical trial research to patients everywhere no matter what their skin tone is and no matter where they live.

Reference

1. Wu D, Lu X, Nakamura M, et al. A pilot study to assess the reliability of digital image-based PASI scores across patient skin tones and provider training levels. Dermatol Ther (Heidelb). 2022;12(7):1685-1695. doi:10.1007/s13555-022-00750-w