IDEA-Based Transvaginal Ultrasound Matched Laparoscopy Closely in Deep Endometriosis
Deep endometriosis often announces itself through pain, distorted pelvic anatomy, and operative findings that can be far more extensive than symptoms suggest. In this prospective pilot study from a tertiary referral centre in India, transvaginal sonography performed with the International Deep Endometriosis Analysis, or IDEA protocol, tracked closely with laparoscopic findings in women already scheduled for surgery. The same scan was used to assign UBESS staging, a system designed to estimate surgical complexity before the first incision.
The cohort was small but clinically pointed: 33 women with suspected deep endometriosis underwent preoperative ultrasound by a single experienced radiologist, then laparoscopy served as the reference standard. Surgical findings split the group into two categories only. Sixteen women had UBESS Stage III, the higher-complexity category, and 17 had Stage II disease. No one in this surgical cohort fell into Stage I.
That absence matters because it narrows the clinical range the test was asked to handle. Even so, the scan performed well where the operative stakes are highest.
Surgical Complexity Was Identified Before the Operation Started
For UBESS Stage III, the ultrasound staging system correctly classified 13 of 16 women, yielding a sensitivity of 81.3% with a 95% confidence interval of 54.4% to 96.0%. Specificity reached 100%, with a 95% confidence interval of 80.5% to 100%, because no Stage II case was incorrectly escalated to Stage III. Overall accuracy was 90.9%, with a 95% confidence interval of 75.7% to 98.1%.
Three Stage III cases were under-staged. In practical terms, those are the cases that can alter theatre planning, operative time, and the threshold for bringing in more experienced surgeons or additional resources. Deep endometriosis behaves less like a superficial implant disease and more like a fibrotic infiltrative process, often tethering pelvic structures together in ways that are hard to appreciate without targeted imaging. When ultrasound undercalls the extent of disease, the operating room learns it late.
The scan did not overcall complexity in any patient. That pattern is reassuring in a preoperative tool, because false escalation can push patients toward unnecessary anxiety or overly aggressive planning, while false reassurance can leave a surgical team underprepared.
Endometriomas and Pouch of Douglas Obliteration Were Detected Reliably
The ultrasound assessment also handled several of the most useful anatomic markers of deep endometriosis. All laparoscopically confirmed endometriomas were identified by transvaginal ultrasound, with right-sided lesions detected in 30 of 30 cases and left-sided lesions in 29 of 29 cases. Sensitivity was therefore extremely high, although the specificity estimates rested on very small numbers of patients without endometriomas, only three to four depending on the side.
For pouch of Douglas obliteration, sensitivity was 85.7% based on 18 of 21 cases, with a 95% confidence interval of 63.7% to 97.0%. Ovarian immobility performed similarly, with sensitivity between 81.3% and 84.2%. No false-positive findings were recorded for these structural markers.
Those are not minor details. Obliteration of the pouch of Douglas and fixation of the ovaries often signal adhesive disease that changes dissection planes and raises surgical difficulty. A scan that can identify those changes before surgery gives the operating team a map of where anatomy has likely been erased by fibrosis.
Why This Imaging Pattern Fits Deep Endometriosis Biology
Deep endometriosis is not simply ectopic endometrial tissue sitting in isolation. It can drive chronic inflammation, scarring, and tethering across pelvic compartments, which is why laparoscopy often reveals disease that is both more extensive and more surgically demanding than clinical examination predicts. The IDEA protocol is built to interrogate exactly those structural consequences: nodules, organ fixation, obliteration of normal spaces, and endometriomas.
UBESS adds another layer by translating the scan into a staging framework tied to operative complexity. In this cohort, that staging system separated Stage II from Stage III disease without overestimating severity, a useful feature for referral pathways and operative preparation. A preoperative test that avoids false escalation while retaining reasonable sensitivity has clear value in a condition where surgery can be prolonged, technically difficult, and dependent on specialist expertise.
Single-Operator Expertise Strengthened the Signal and Limited the Reach
The study design strengthens confidence in the measurements, but it also defines the ceiling of what can be inferred. A single experienced radiologist performed all examinations, which reduces interobserver variability but also makes the results closely tied to specialist skill. This is the kind of performance that can be difficult to reproduce in less specialized settings.
The sample size was small, and the confidence intervals were wide. Absence of Stage I disease also means the test was not challenged across the full spectrum of endometriosis severity. In addition, the specificity estimates for endometriomas were based on very few negative patients, so those numbers should not be treated as stable population estimates.
Still, the operative reference standard was direct and clinically relevant. Laparoscopy remains the benchmark for confirming deep endometriosis, but it is invasive and only arrives after the decision to operate has already been made. Imaging that can sort high-complexity cases from less demanding ones before surgery can influence counselling, operative scheduling, and the choice of surgical team.
A Small Cohort With Clear Preoperative Consequences
The most actionable finding is straightforward: IDEA-based transvaginal ultrasound identified endometriomas and pouch of Douglas obliteration with few errors, and UBESS missed three of 16 high-complexity cases without over-staging any woman. For a disease in which anatomy often dictates operative risk, that is a clinically meaningful pattern.
The unresolved question is not whether the approach can work in expert hands. It is whether the same accuracy will hold across larger multicentre cohorts, with broader disease severity and a wider mix of operators. Until that is tested, the findings support careful optimism rather than routine generalization.