Welcome back to the journal review. This is JAMA Dermatology, February 2026, and we've got two pieces worth your time this episode — a viewpoint on something every one of us does dozens of times a day without thinking about, and a research letter that should make you a little uncomfortable about how well your patients actually understand what you're telling them.

Let's start with the viewpoint, titled "Effects of Photographic Image Processing in Dermatology," from Gall, Wooten, and Lyford out of Naval Medical Center San Diego. This is an opinion piece, not a study, so there's no methods or results section to walk through — instead it's an argument, built on cited background and the authors' own reasoning, about a blind spot in how we use clinical photography.

The setup is simple: about three-quarters of dermatologists now regularly use their own smartphone to photograph patients in clinic. That convenience comes at a cost we rarely think about. Modern smartphones don't just capture an image — they process it. Multiple exposures get layered together in what's called high dynamic range imaging, edges get sharpened, colors get warmed and saturated, all automatically, all before you ever see the photo. The authors' point is that this is a double-edged sword in dermatology specifically. That same processing that makes a landscape photo look nicer can smooth out subtle textural change, or conversely make erythema or pigmentation look more dramatic than it really is. Either direction is a problem when you're trying to make a call about lesion evolution or a subtle inflammatory finding based on that photo six months later.

Then there's a second, separate insult that happens after the photo is taken well: the electronic health record itself. Most EHR platforms apply lossy compression on upload — stripping metadata, cutting resolution — to save storage and speed up loading. The authors are explicit that this degradation is irreversible; once that data is discarded, you cannot get it back. So you can have a technically excellent original photograph that becomes diagnostically inferior the moment it lives in the chart.

The authors illustrate this directly with a side-by-side figure — the same rash photographed with standard automatic smartphone processing, then as an unprocessed RAW file from a third-party app, then again after that standard image was uploaded into a commercial EHR. The visual differences between those three versions are the entire argument in one picture.

Their practical recommendations are worth internalizing even though this is an opinion piece rather than a trial. If you want an unaltered image, RAW-capture apps exist, but they need to be HIPAA compliant, and remember a RAW file requires better underlying photographic technique since none of the automatic corrections are there to bail you out. When uploading to a computer or the EHR, make sure your phone is set to transfer full-resolution files, not thumbnails, avoid email as a transfer method since it's particularly prone to degrading images further, and ideally maintain a separate unadulterated archival copy in a low-compression format like TIFF. Their most concrete institutional suggestion is a disclaimer within the EHR itself, flagging whether an image has undergone automatic smartphone editing, paired with discouraging use of the uploaded, potentially degraded EHR image for actual diagnostic or treatment decision-making.

Is this practice-changing today? Not in the sense of new data forcing your hand — the authors themselves note there's no existing research quantifying how much these alterations actually affect diagnostic accuracy or outcomes. But it is a legitimate awareness-changing piece. For those of us who use serial photography for margin mapping, recurrence surveillance, or field cancerization monitoring, the takeaway is to treat your EHR-embedded photo as a reference thumbnail, not a diagnostic-grade document, and to maintain your own high-fidelity image archive outside that pipeline if precise comparison over time actually matters to your management.

Now to the second piece, a research letter titled "Patient Comprehension of Skin Cancer–Related Terminology," out of Beth Israel Deaconess and Brigham and Women's.

This is a survey study, and a fairly tightly scoped one. The background problem is straightforward — good communication drives outcomes, and dermatology in particular leans on a dense vocabulary of oncology-specific terms that we use fluently and reflexively. The gap here is that nobody had systematically quantified how much of that vocabulary patients actually understand.

Methodologically, the authors ran an anonymous, voluntary paper survey across general dermatology clinics over a two-week window in May 2025, testing comprehension of thirteen common skin cancer-related terms via multiple-choice questions with a single correct answer, then used multivariable linear regression adjusted for demographics to look at what predicted the mean percent-correct score. A survey is really the only tool that fits this question — you cannot infer comprehension from a chart, and an interventional design would be premature before you even know the baseline gap exists. Excluding patients needing interpreters or presenting for cosmetic visits was a sensible way to isolate the population where oncologic terminology actually matters clinically.

They got a strong response — 166 of 182 eligible patients completed it, over 90%. Overall mean accuracy across all thirteen terms was about 69%, which sounds passable until you look at which terms were driving that number down. Comprehension of straightforward words like topical or skin biopsy was excellent, up around 95%. But understanding cratered for two terms you and I use constantly: only about 14% of patients correctly defined actinic, and only about 22% correctly defined dysplastic nevus. Clear margins and pathology results — phrases we deploy in nearly every postoperative conversation — were understood by only around six in ten patients, meaning roughly a third of patients walking out of a visit where you said "we got clear margins" may not actually know what that means.

Perhaps the most clinically sobering figure: not everyone recognized the cancer diagnoses themselves as cancer. About 82% correctly identified melanoma as a skin cancer, but that dropped to around 70% for basal cell carcinoma and under 60% for squamous cell carcinoma — meaning a meaningful minority of patients being told they have a basal cell or squamous cell carcinoma may not register that they've just been given a cancer diagnosis at all.

On subgroup analysis, younger patients age 18 to 29 scored significantly and substantially lower than those 60 and older — a roughly 18-point gap, which is both statistically significant and clinically meaningful, somewhat counterintuitive if you assume younger patients are more health-literate by default. Male patients scored significantly lower than female patients. Graduate education and more than five prior dermatology visits both tracked with higher scores, which is expected, but even in those higher-performing groups, average accuracy stayed under 75% — the authors make the point explicitly that routine exposure to dermatologic care is not, by itself, enough to guarantee real comprehension.

Limitations are honestly stated and matter for how far you generalize this: single-site, mostly White patient population with too few Asian, Black, multiracial, and Hispanic or Latino respondents to draw race-specific conclusions, no data collected on family history or the reason for the visit, and a likely overestimate of true population comprehension given the Boston area's overrepresentation of health care professionals among respondents.

Practically, this one is directly actionable in your own consult room, even without further trials. It's not practice-changing in the sense of altering a surgical or diagnostic protocol, but it is a strong, well-supported nudge to change your language habits. If a third of patients don't reliably know what "clear margins" or "pathology results" means, and a large minority don't register squamous cell or basal cell carcinoma as cancer, that has real implications for informed consent conversations, adherence to surveillance recommendations, and how you frame Mohs results at the post-op visit. Simple fix, no cost, immediate implementation: replace jargon with plain-language equivalents in the moment, and consider explicitly confirming comprehension of terms like "margins" and "dysplastic" rather than assuming they land the way you intend.

That's the episode — a reminder that the images we rely on and the words we use both carry more hidden distortion than we tend to assume. Thanks for listening, and we'll see you next month.