Welcome back to the journal review. This is the May twenty twenty-six issue of JAMA Dermatology, and we've got two original investigations worth your time this episode — a baseline characteristics paper from a practice-changing surveillance trial, and a systematic review and meta-analysis that sharpens the dermoscopic criteria you're already using at the bench. Let's get into it.

First up is a paper titled "Characteristics of Participants Screened and Randomized to the Melanoma Self Surveillance Trial." This is best described as a secondary baseline analysis nested within a large randomized controlled trial — essentially the "who showed up" paper that precedes the outcomes paper. It's not reporting efficacy yet; think of it as the demographic and psychological scene-setting for the main event still to come.

The clinical problem here is one you live with every week: after excision of early-stage melanoma, there's no consensus on how often to bring patients back for scheduled surveillance visits, and there's no solid comparative evidence that more frequent scheduled visits actually catch more melanomas or improve survival. Meanwhile, we know a substantial fraction of new primaries and recurrences are actually caught by patients or their partners between visits, not at the scheduled appointment itself. That's the rationale behind patient-led surveillance — arming patients with structured skin self-exam support, a clip-on dermatoscope, and teledermatology backup, with a fast-tracked visit if something looks wrong, rather than just waiting for the next calendar slot.

This is the MEL-SELF trial, a two-arm, multicenter randomized trial run across specialist dermatology and GP-led skin cancer clinics in Australia, enrolling patients with AJCC stage zero through two melanoma. What's methodologically interesting, and worth flagging for anyone reading trial design sections, is the active run-in phase before randomization — participants had to actually complete the baseline questionnaire, watch structured self-exam instructional videos, and successfully upload photos of a target lesion before they were allowed into the randomization pool. The authors' stated rationale is that this weeds out people who can't or won't adhere to the intervention tasks, which protects the internal validity of the eventual efficacy comparison — you don't want your trial diluted by people who were never going to follow through with the app regardless of arm. They also used minimization rather than simple randomization, balancing groups dynamically on stage, age, sex, clinic type, calculated melanoma risk, and dysplastic nevus status — a sensible choice for a moderately sized trial with several prognostically important variables where simple randomization could easily leave you with unlucky imbalance.

Now, this particular paper isn't reporting the trial's primary outcome — that comparative analysis of detection rates is still to come in a future publication. What we get here is the baseline snapshot, and honestly, it's clinically revealing in its own right. Of about twelve hundred patients screened as potentially eligible, five hundred and four were randomized, split roughly evenly across the two arms. The population skews slightly female, mean age mid-fifties, and importantly nearly all had lower-risk disease — melanoma in situ or stage IA made up roughly nine in ten participants. That matters for how you interpret any eventual finding about missed lesions; this is not a high-stage cohort.

The most clinically actionable numbers are about baseline behavior and psychology, not about the intervention itself. Skin self-exam practice was frankly suboptimal — one in five participants had done no self-exam at all in the preceding year, and about half lacked confidence in doing it properly, even though this is a population that by definition has already had melanoma and is under active surveillance. That's a real unmet-need finding independent of trial outcome. Equally notable: nearly half of participants met criteria for clinically significant fear of cancer recurrence. And here's the nuance worth remembering when you're counseling patients — that fear correlated with being female, younger age, and higher depression and anxiety scores, and it correlated with a patient's perceived lifetime risk of a new melanoma, but it did not correlate at all with their actual calculated statistical risk of a new primary. In other words, fear of recurrence in this population is driven by psychology, not by objective risk — the odds ratio linking fear to true calculated risk was essentially flat, right around one. That's a statistically clean null finding, and it's clinically meaningful: it tells us that reassurance based on objective risk numbers alone is unlikely to move the needle on a patient's fear, because their fear isn't anchored to that number in the first place.

The authors are appropriately restrained in their conclusions here — this is descriptive, hypothesis-generating baseline data, not an efficacy claim. The obvious limitation is that this is a highly select, motivated, largely low-stage Australian cohort that survived an active run-in filter, so generalizability to less digitally engaged or higher-stage populations is uncertain; the authors do note that characteristics were similar between randomized participants and the potentially eligible-but-not-randomized group, which is somewhat reassuring against selection bias, but it's still not a general surveillance population.

So what do you do with this today versus what you file away? Nothing here is practice-changing yet — the actual comparative outcome data on whether patient-led surveillance changes detection or psychological outcomes hasn't been published. What is immediately useful is the reminder, backed by real numbers, that a substantial minority of your post-melanoma patients are essentially not doing meaningful self-surveillance between visits, and that fear of recurrence is common and decoupled from actual risk — which argues for structured self-exam coaching and psychological screening as part of routine surveillance visits regardless of how the main trial reads out.

Now to the second article, "Diagnostic Accuracy of Dermoscopic Features in Acral Lentiginous Melanoma: A Systematic Review and Meta-Analysis." This is exactly what it says — a PRISMA-adherent systematic review with meta-analytic pooling, addressing a problem that's squarely in your daily differential: distinguishing acral lentiginous melanoma from benign acral nevi dermoscopically, particularly relevant given ALM's disproportionate representation in patients with skin of color and its generally poor prognosis at advanced stage.

The gap the authors identify is that while the parallel ridge pattern has long been taught as the specific marker for ALM, and parallel furrow as the benign counterpart, that binary framework leaves real diagnostic uncertainty — plenty of ALMs don't show a classic ridge pattern, and relying on just two features undersells the more nuanced pattern recognition dermoscopists actually use.

Methodologically, this is a straightforward and well-executed synthesis — five databases searched from 1970 through early 2023, independent dual screening and extraction with adjudication, and formal risk-of-bias assessment tailored to study design: Newcastle-Ottawa for cohort studies, QUADAS-2 for diagnostic accuracy studies, and the Joanna Briggs Institute checklist for cross-sectional studies. That tripartite quality assessment approach makes sense here because the 41 included studies weren't a single uniform design — pooling proportions across heterogeneous study types requires matching the appraisal tool to what was actually being measured in each. They used random-effects models for pooling, which is the right call given the expected clinical and methodological heterogeneity across decades of acral lesion studies from different populations, and they checked publication bias with Deeks' funnel plots, appropriate for diagnostic proportion data.

The final pooled dataset comprised close to nine thousand acral nevi from thirty-five studies and about eight hundred ALMs from eighteen studies. The headline result confirms what you already practice: the parallel ridge pattern remains the standout discriminator, present in roughly four out of five ALMs but essentially absent in nevi — under one percent. That's a pooled difference of about seventy-seven percentage points, and it's both statistically rock-solid and clinically about as strong a dermoscopic signal as exists in this field. On the benign side, parallel furrow was the mirror image, seen in about half of nevi but under one in ten ALMs — again a large, clinically meaningful, statistically significant separation.

Where this paper adds genuine value beyond restating the ridge-furrow dogma is in two additional features. Multicomponent structures — meaning multiple distinct dermoscopic patterns coexisting within one lesion — were significantly more common in ALM, roughly half of melanomas versus about one in twenty nevi, though the confidence interval on that estimate was wide, reflecting real heterogeneity across studies, so treat that magnitude as a directionally strong but less precise signal compared with the ridge finding. And latticelike pattern, previously bucketed loosely with benign furrow-type features, was confirmed as significantly more common in nevi than melanoma — about one in eight nevi versus roughly one in forty ALMs — giving it standing as its own independent benign indicator rather than just a variant of parallel furrow.

The other numbers worth knowing for calibration: fibrillar, homogeneous, globular, reticular, and the catch-all "other" and "nontypical" categories were all inconsistent or non-significant across studies — meaning don't lean on those patterns as decision points, the literature simply doesn't support them as discriminators one way or the other.

The authors are direct about limitations, which is appropriate for a review pooling half a century of literature: quality assessment showed retrospective design and inadequate confounder adjustment among the cohort studies, and diagnostic accuracy studies had bias concerns around nonrandom patient selection and inconsistent histopathologic confirmation of control nevi — meaning some "benign" comparator lesions weren't always biopsy-proven, which could inflate apparent specificity of malignant features. Funnel plot asymmetry was also flagged for some features, particularly fibrillar and multicomponent patterns in melanoma studies and furrow, latticelike, and nontypical patterns in nevi studies, raising some possibility of small-study or reporting bias inflating those specific effect estimates. And clinically, the authors are honest that this doesn't solve the whole problem: close to one in five ALMs lacked the parallel ridge pattern altogether, meaning a meaningful minority of these lesions will not show the textbook finding, and multicomponent or latticelike features don't fully close that gap either.

Practically, here's the takeaway distinction. The confirmation that parallel ridge and parallel furrow remain your primary anchors is not new — that's validation, not change. What is genuinely useful and worth incorporating into how you teach or how you dermoscopically document equivocal acral lesions is elevating multicomponent structure as a real red flag deserving biopsy consideration even in the absence of a ridge pattern, and recognizing latticelike pattern as an independently reassuring benign feature rather than lumping it vaguely with furrow. Neither of these should change your threshold for biopsying a clinically concerning acral lesion — dermoscopy remains an adjunct, not a replacement for histopathology, and the persistent one-in-five ridge-negative ALM rate is the number to keep front of mind when a lesion looks off clinically despite reassuring dermoscopy.

That wraps this episode. Bottom line: the MEL-SELF baseline data tells us our post-melanoma patients are under-practicing self-surveillance and carrying fear disconnected from actual risk, with the real efficacy verdict still pending — worth watching, not yet actionable. And the acral melanoma meta-analysis reinforces your existing ridge-and-furrow framework while giving you two additional, evidence-backed pattern cues — multicomponent and latticelike — to sharpen your threshold for biopsy on ambiguous acral lesions. Thanks for listening, and we'll see you next month.