Review Papers Against Their Own Results

 Quick Explanation Copied

 Long Explanation

Historical Anchor: Why This 1998 Review Still Shapes Practice

Reinhold, Tafazoli, and Wang (McGill/Montreal General Hospital) synthesized the pre-operative imaging evidence for adenomyosis at a moment when high-resolution EVS and MRI had just transformed a formerly surgical diagnosis into an imaging diagnosis Clinical pre-op diagnosis was 2.6-26%. The review's core mechanistic insight — that smooth muscle hyperplasia produces the hypoechoic/low-signal lesion, while heterotopic basal endometrium explains echogenic islands, myometrial cysts, and high-signal foci — mapped histology onto imaging signs in a way that remains the teaching standard. The authors' own prospective hysterectomy cohort is the evidentiary backbone: among 100 consecutive women (29 with adenomyosis), EVS achieved 86% sensitivity (95% CI 74-98%), 86% specificity, 94% NPV, and 71% PPV, with myometrial cysts in ~50% of cases Authors. They also introduced the practical JZ dichotomy (≥12 mm highly predictive; ≤8 mm usually excluding), grounded in their 119-patient MRI series where mean JZ was 7.7±3.3 mm without and 15±4.9 mm with adenomyosis JZ thickness means from MRI series.

Modality-by-Modality Verdict (as reported, then stress-tested)

• HSG: correctly retired — the classic spicule sign appeared in only 25% (38/150) of proven cases, and myometrial contrast pools are non-specific (vascular/lymphatic extravasation mimics) HSG spicule sign in only 25% of cases.
• TAS: insufficient resolution; Siedler 1987 reported sensitivity 63%, specificity 97% — unusable alone for treatment decisions.
• EVS vs MRI: the authors found no significant sensitivity/specificity difference in their 119-patient head-to-head (EVS sensitivity 89% vs 53% in Ascher's earlier series, a discrepancy they attribute to recruitment bias, hard-copy review, and higher focal-disease prevalence) EVS-MRI head-to-head results and discrepancy.

Note: Specificity bars for Reinhold MRI and Ascher EVS reflect values reported in or inferable from the review text; Ascher EVS specificity (97%) is the reported NPV-adjacent figure for leiomyoma exclusion — treat with caution, as the review does not tabulate all specificities uniformly. This uneven reporting is itself a critique: the pre-PRISMA literature it synthesized lacks standardized accuracy definitions Heterogeneity and bias in imaging literature.

Blind Spots, Later Refutations, and What Would Change the Verdict

• JZ threshold fragility: the ≤8 mm exclusion rule was challenged by Kang et al. (cited within the review itself), who found 40% of normal volunteers had regional JZ >5 mm — the review honestly flags this, a credit to its skepticism.
• False positives mimicking adenomyosis: the review warns of muscular hypertrophy and transient myometrial contractions, but it could not anticipate uterine metastasis from lobular breast cancer producing a 20 mm JZ with high-signal foci — a 2004 case report showing the MRI criteria can be malignant mimics Uterine metastasis mimicking adenomyosis on MRI.
• Later head-to-head data partially reversed the EVS-MRI equivalence: in 106 premenopausal hysterectomy patients, MRI specificity (86%) exceeded TVS (65%) with similar sensitivity (~68-70%) — the opposite emphasis of Reinhold's conclusion that EVS should be first-line 2001 head-to-head MRI vs TVS accuracy, and the 2018 meta-analysis similarly favored MRI (pooled sensitivity 77%, specificity 89%) Pooled MRI vs TVS meta-analytic accuracy. Cost and availability still justify EVS-first triage in practice.
• Unanswered questions the review opened: the histological basis of JZ low signal was explicitly unsettled; the clinical relevance of imaging-detected disease (up to 70% prevalence in unselected hysterectomy specimens) was acknowledged but unresolved, and subsequent work showed ultrasound-diagnosed adenomyosis halves viable IVF pregnancy rates (23.6% vs 44.6%) Adenomyosis halves viable IVF pregnancy rate, and EHR modeling suggests ~19 per 1000 potentially undiagnosed cases versus 1.7 diagnosed — a ~10-fold under-ascertainment gap that imaging was supposed to close EHR-based adenomyosis prevalence and underdiagnosis.

Confidence and falsifiability: The review's quantitative claims rest on its own prospective cohorts plus small single-center series — moderate external validity. The conclusion that EVS should be first-line would be falsified by consistent multi-center superiority of MRI at acceptable cost, which the 2001 and 2018 comparisons approximate but do not definitively establish given operator-dependence and classification heterogeneity. No placebo/treatment-arm issue applies; the notable funding/conflict gap (none declared in 1998) contrasts with later industry-funded prevalence work AbbVie funding and conflict disclosure.

Best Evidence: Adenomyosis Imaging TodayGraph Pooled Diagnostic Accuracy

Author Review: Caroline ReinholdAuthor Review: Faranak TafazoliAuthor Review: Lin Wang

 BGPT Paper Review

Study Novelty

In 1998, consolidating EVS/MRI criteria and introducing a usable JZ thickness dichotomy was genuinely novel for a condition previously diagnosed only at hysterectomy; the individual findings were pre-existing but the histologic-imaging mapping and thresholds were new synthesis.

Scientific Quality

Grounded in the authors' own prospective 100- and 119-patient hysterectomy cohorts with histopathologic gold standard and reported confidence intervals; weaknesses include uneven accuracy reporting across cited studies, small series, and no formal systematic search.

Study Generality

Applies broadly to all gynecological imaging practice and differential diagnosis with leiomyoma and endometrial carcinoma, not a niche technical question.

Study Usefulness

Directly changed clinical pathways: retired HSG, established EVS-first triage with MRI for indeterminate cases, and provided criteria (JZ thresholds, cyst size, adenomyoma vs leiomyoma features) used in teaching and guidelines for decades.

Study Reproducibility

Prospective design and explicit numeric criteria are reproducible in principle, but real-time EVS diagnosis is operator-dependent and cannot be replicated from hard copy — a limitation the authors themselves acknowledge.

Explanatory Depth

Offers a strong histopathologic mechanism (muscle hyperplasia vs heterotopic basal endometrium) explaining each imaging sign, but leaves the JZ low-signal basis and clinical relevance of mild disease explicitly unresolved.