Postoperative Cysts: What They Actually Mean

→ Small cysts (fluid-filled holes in bone) often show up on MRI after cartilage repair. They look alarming, but new research shows they don't necessarily mean the surgery failed. Don't panic at the scan.

Added August 6, 2026 — what an untreated cystic lesion actually does over eight years, and this site was arguing about the paper without holding it

→ This is the single most relevant new finding of the week, and it is reassuring. Somebody followed 56 people who had exactly this — a talar lesion with a subchondral cyst — and did not operate on them, for a median of eight years. The cysts mostly did nothing.

How it was found is worth stating. This site already cited a letter criticising this study [221] and the authors' response to that letter [222] — but never the study itself. It was hosting a debate about a paper it had not read.

Seo JH, Choi WS, Shin MY et al. — “Radiologic Progression of Talar Osteochondral Lesions With Subchondral Cyst: A Median 98-Month Observational Study,” Foot & Ankle International 2026 [389]. 56 patients with cystic OLT managed non-operatively, median follow-up 98 months — over eight years.

• Most cystic lesions left alone simply do not move. Median annual cyst volume change across the whole cohort was 0.3 mm³/year, and roughly three-quarters of patients either grew by less than 8 mm³/year or actually shrank
• But a quarter behave completely differently — the upper quartile enlarged at a mean of about 137 mm³/year. This is a bimodal problem, not a uniformly benign one, and the whole question is which group he is in
• Three predictors separate the two groups, and two of them are readable off imaging he already has: larger initial cyst volume (OR 1.002, P = .03), bone marrow edema (OR 10.19, 95% CI 1.68–61.97, P = .012), and wall breakage — loss of subchondral bony integrity — (OR 16.94, 95% CI 2.73–105.35, P = .002)
• Cyst progression had no correlation with symptoms (P > .05), and “none of the cases progressed to OA” over the median eight years

→ The practical consequence, and it is a specific request rather than a feeling: ask whoever next reads his MRI two questions — is there bone marrow edema, and is the cyst wall broken? Those two answers carry odds ratios of 10 and 17 for progression, and they cost nothing to obtain because the images already exist. That is a far better-grounded way to decide how urgent this is than watching the pain and guessing.

Read with its real limits. This is a retrospective observational cohort, not a trial, and the abstract was read in full but the full text was not retrieved. More importantly, these are primary lesions under observation, not a lesion under a failed structural allograft — a graft-host interface that has already failed once is a different mechanical situation from a cyst that has never been operated on. And the Amsterdam group's published letter [221] contests the paper's methods, which is precisely why holding all three documents together is better than holding two of them.

Added August 8, 2026 — a randomized trial where the grafts with more small cysts did better, which is this page's whole thesis wearing a Level I badge

→ A blinded trial compared allografts soaked in concentrated bone marrow against plain ones. The soaked grafts showed more small cysts on CT — and needed reoperation seven times less often. If small cysts on a scan reliably meant failure, that result would be impossible.

Yanke et al., Rush, double-blind RCT, n=36 knee osteochondral allografts — BMAC-treated grafts were “more likely to have small cystic changes (P = .01), with an associated trend toward reduction in large cyst formation (P = .06),” yet underwent reoperation at 5.3% versus 35.3% for the untreated grafts (P = .02) [405]. The authors' own reading is that small cystic change may be remodeling activity rather than failure — the same direction as the meta-analysis below and the Seo natural-history cohort above, but from a randomized, blinded design. The distinction it sharpens is small versus large: small cystic change coexisted with the best outcomes; what the BMAC arm trended away from, and what the Seo predictors (wall breakage, edema) flag, is the large, progressing cyst. Knee, not talus; n=36; no PROM differences — but the next time a follow-up scan of this ankle shows a small cyst, this is the trial to hold onto before drawing conclusions.

Added August 13, 2026 — the ordinary MRI sequences on every follow-up scan were finally tested against ground truth, by the people who wrote the scoring system

→ When a radiologist says the bone under a repair “looks brighter” or “looks darker,” is that actually measuring healing — or just describing pixels? A Vienna group implanted scaffolds in 28 sheep, scanned them with the same routine sequences used on human follow-up MRIs, then cut the joints open and checked the answer under a microscope and micro-CT.

The Vienna High-Field MR Center — senior author Trattnig, originator of the MOCART score used across this site’s imaging entries — validated routine T1-SE and PD-FSE signal readings against micro-CT bone volume and histology in an ovine osteochondral-repair model at 30, 180 and 365 days [416]. Blinded radiologist ratings on both sequences tracked real new-bone formation; PD-FSE contrast-to-noise fell as regeneration progressed (P < .001) — the repair zone becoming less conspicuous is the healing signal — while T1-SE contrast-to-noise alone did not track mineralization and, the authors state, should not be used as a stand-alone marker. Inter-reader agreement was excellent (ICC ≥ 0.947).

Why it belongs on this page: the running thesis here is that post-operative scans are systematically over-read — cysts, edema and signal change get treated as verdicts when they are often remodeling. This is the first study to anchor the routine sequences themselves to ground truth, and it cuts both ways: some ordinary readings genuinely do index subchondral healing (so a follow-up MRI is not theater), and one common quantitative shortcut genuinely does not. Limits: sheep knees, scaffold repair, not human tali or allografts — it informs how a follow-up scan is read, not what to do about it.

2025 Meta-Analysis

A 2025 meta-analysis examined postoperative cyst rates and their relationship to clinical outcomes across different graft types [150]:

• Did not find a significant relationship between cyst presence and clinical outcomes
• Small cysts alone do not automatically mean the joint-preserving strategy has failed
• Osteoperiosteal grafts had the lowest postoperative cyst rate

Clinical significance: This matters because imaging-based anxiety can push patients toward unnecessary revision. Postoperative cysts should be interpreted alongside symptoms, not in isolation.