D48.5 or D49.2? The one question that decides it

D48.5 pays. A lesion gets biopsied, the diagnosis isn't confirmed, the coder reaches for "neoplasm of uncertain behavior of skin," and the claim clears. That reflex is why it became the default in so many practices, and why it shouldn't be the go-to code. It pays, but most times it's used it's still the wrong call. Neoplasm-behavior coding has a standard with a specific right answer, and the good news is that answer is usually one your MAC covers too.

It's not uncommon to see providers run almost everything through D48.5: biopsies, shaves, and excisions, including lesions that were already biopsied and read. That is the archetype: D48.5 as the catch-all for "we took something off" rather than a code earned by a finding. It's the same biopsy-and-pathology workflow that already bills three ways off one specimen; the behavior code is one more place the standard slips.

D48.5 is "Neoplasm of uncertain behavior of skin." That word sounds like it describes your state of mind while you wait on the lab. It doesn't. "Uncertain behavior" is a histologic conclusion: the specimen went under the microscope and the pathologist still couldn't call it benign or malignant, whether from atypia, borderline features, or "cannot exclude melanoma." Before the path report is back, you don't have uncertain behavior. You have a lesion you haven't characterized yet, and the code has to match what you know when the visit ends. That's less than D48.5 claims.

Nothing at billing pushes back on it. Payers don't pull the pathology report at claim time; adjudication runs automated diagnosis-to-procedure and coverage edits against the fields you submit, nothing more. The record gets read later, on a post-payment review, where payers score the documentation against the code the same way they scrutinize a level-four E/M. So a code that overstates the record clears, gets paid, and the practice never sees a reason to change it. The pull is systematic, not random: favorites lists surface top-of-alphabet codes first, "suspicious lesion" templates equate clinical suspicion with documented uncertain behavior, and carry-forward keeps D48.5 on the problem list after path has resolved it. The workflow is picking the code, not the coder.

The two codes aren't interchangeable defaults but mutually exclusive states, and the code book says so. D48.5 lives in the "uncertain behavior" range (D37 through D48), where the block note says histologic confirmation of malignant versus benign cannot be made. D49.2 is "Neoplasm of unspecified behavior of bone, soft tissue, and skin," and its range carries a reciprocal Excludes1 against the uncertain codes. Each block excludes the other, so you can never correctly code both for the same lesion: one says the pathologist looked and couldn't decide, the other that nobody has recorded the behavior at all. At a pre-path biopsy visit, only the second is usually true.

The outpatient guideline points the same way. ICD-10-CM Official Guidelines Section IV.H says not to code "probable," "suspected," "questionable," "rule out," "compatible with," or "working diagnosis," but to code "to the highest degree of certainty for that encounter/visit, such as symptoms, signs, abnormal test results, or other reason for the visit." Derm offices are outpatient, so that rule governs.

At the biopsy encounter, before path returns, run the note through three questions. If the note describes only a lesion, a changing spot, or abnormal skin and never calls it a neoplasm, code the lesion or sign, often L98.9. That's the honest ceiling for most pre-path notes. When the provider documents a neoplasm, tumor, or growth but states no behavior, D49.2 becomes defensible, though only because the chart actually says neoplasm. And unless the provider has documented indeterminate behavior outright, D48.5 stays off the claim.

Doing it right adds one step that belongs to the standard, not outside it as a warning. The correct code is the one that is both supported by your documentation and on your MAC's covered-diagnosis list. Coverage isn't symmetric: some MAC articles list both D48.5 and D49.2 as supporting medical necessity, others list D48.5 and leave D49.2 off. So the standard isn't "swap to D49.2." It's code what the documentation supports, check your MAC's covered list, hold for path where the workflow allows, and re-code once the report lands.

When path returns, code to the finding: benign nevus to D22.x, other benign to D23.x, non-melanoma skin cancer to C44.x, melanoma to C43.x or D03.x in situ, and genuinely indeterminate reads to D48.5, where it's finally correct. Fix an already-submitted claim with a frequency-7 corrected claim, not a void. The clinician's pre-path code and the pathologist's later read are two separate claims that aren't supposed to match. That divergence is the normal sequence, not a mismatch waiting to deny.

A behavior code the chart doesn't support pays clean and then reverses later, usually as a simple recoupment, and in a bad audit sample as an extrapolated one projected across more claims than the few that were actually miscoded. It's the same principle behind the documentation defect that quietly denies flap claims: the note has to earn the code. CMS's own FY2025 CERT data puts the Medicare fee-for-service improper-payment rate at 6.55%, roughly $28.83 billion, with insufficient documentation named as the single largest root cause. That is an argument for coding to the standard, not a reason to lose sleep over any one claim.

Clarity RCM manages revenue cycle for 200+ dermatology practices across 42 states. It's all we do. See how we work.