Can a Pelvic Ultrasound Detect Endometriosis?

A pelvic ultrasound can detect endometriosis, but how much it catches depends on the type of disease, where it has grown, and who is performing the scan. Ovarian endometriomas and deep deposits in the bowel tend to show up well, while shallow implants scattered across the pelvic lining are frequently missed. The difference in detection rates between these forms of the disease is dramatic, and the experience of the person holding the probe can matter as much as the technology itself.

Ovarian Endometriomas Are the Easiest to Spot

If endometriosis has formed a cyst on the ovary, often called a “chocolate cyst” because of the dark old blood inside, transvaginal ultrasound picks it up reliably. These endometriomas have a characteristic appearance on the screen: low-level internal echoes that look like ground glass, well-defined walls, and usually fewer than five internal compartments. Across multiple analyses, the sensitivity and specificity of ultrasound for ovarian endometriomas are both in the range of roughly 70 to 96 percent, depending on the study and the operator’s skill, and these figures are comparable to what MRI achieves for the same lesions.¹ Research into texture-analysis software has shown that computer-assisted evaluation of the ultrasound image can push accuracy even higher, outperforming visual assessment alone when distinguishing an endometrioma from an ordinary hemorrhagic cyst that can look similar.²

In practical terms, this means a standard pelvic ultrasound ordered by your doctor is well suited to finding ovarian endometriomas. Many cases of endometriosis are first suspected on exactly this basis: a cyst appears on a routine scan, and its appearance raises a red flag. The trouble is that endometriosis rarely confines itself to the ovary. And for the disease that lives outside the ovary, the picture gets more complicated.

Deep Infiltrating Endometriosis and Where It Hides

Deep infiltrating endometriosis grows more than five millimeters below the peritoneal surface and can invade the bowel wall, the bladder, the ligaments that support the uterus, the rectovaginal space, and the vaginal wall itself. Transvaginal ultrasound can find many of these deposits, but sensitivity swings widely depending on the anatomical site. A systematic review found that for rectosigmoid involvement, ultrasound achieved pooled sensitivity of about 91 percent and specificity of 97 percent, making it an excellent tool for bowel disease.³ Another systematic review reported overall deep-endometriosis sensitivities between roughly 79 and 85 percent and specificities reaching 92 percent.⁴

Site-specific numbers tell the real story, though. Bladder endometriosis is detected with a pooled sensitivity around 62 to 72 percent and near-perfect specificity.⁵ The uterosacral ligaments are harder: sensitivity for disease there has ranged from as low as 10 percent to as high as about 89 percent across different studies, depending heavily on operator expertise and technique.⁴ The rectovaginal septum falls somewhere in between, with pooled sensitivity around 49 to 59 percent.⁵ If you are wondering why so many people with endometriosis hear “your ultrasound looks normal” despite having significant disease, the answer often lies in this patchwork of site-by-site detection rates.

Superficial Endometriosis Is the Biggest Blind Spot

Superficial endometriosis consists of shallow implants that sit on the surface of pelvic organs and the peritoneal lining. It is the most common form of the disease, and unfortunately the hardest to see on any imaging. A large diagnostic-accuracy study using the international consensus protocol found that transvaginal ultrasound had a sensitivity for superficial endometriosis of only about 4 to 44 percent, even though specificity was extremely high, in the range of 99 to 100 percent.⁶ That means when a sonographer does see superficial disease, they are almost certainly right. But the scan misses the majority of cases.

Researchers have been trying to characterize what superficial lesions look like on ultrasound. They can appear as solitary or clustered spots, sometimes protruding from the peritoneal surface, sometimes looking like small dents in it, and they may contain bright spots or wispy adhesions.⁷ One study focusing specifically on the pouch of Douglas peritoneum detected superficial endometriosis with a sensitivity of about 52 percent and specificity of 94 percent, suggesting that targeted examination of specific areas can improve detection somewhat.⁸ But these remain early-stage findings, and for now, a normal-looking ultrasound does not rule out superficial endometriosis. This is the single most important takeaway for anyone researching this question: a “clear” scan does not mean you are disease-free.

The Sliding Sign and Other Dynamic Techniques

One of the most valuable things ultrasound can do for endometriosis actually has nothing to do with seeing a lesion directly. The “sliding sign” is a real-time technique where the sonographer gently presses the transvaginal probe against the cervix and watches whether the front wall of the rectum glides smoothly over the back of the cervix and uterus. If it does, the pouch of Douglas is considered open. If the tissues are stuck together and refuse to slide, the pouch is likely obliterated by endometriosis-related adhesions.⁹

The numbers for this technique are impressive. In a study comparing the sliding sign to what was found at laparoscopy, the technique predicted pouch of Douglas obliteration with an accuracy of 93 percent, sensitivity around 83 percent, and specificity of 97 percent.¹⁰ A negative sliding sign also predicted deep rectal endometriosis itself with 85 percent sensitivity and 96 percent specificity.¹¹ The sliding sign is a good example of why the phrase “pelvic ultrasound” can mean vastly different things. A quick scan that only looks for cysts will not include this maneuver. A dedicated endometriosis scan will.

Why the Operator Matters So Much

No discussion of ultrasound accuracy for endometriosis is complete without addressing the enormous gap between routine scanning and expert scanning. In one retrospective study, routine transvaginal ultrasound detected endometriosis in only 25 percent of surgically confirmed cases, compared to 78 percent for expert-guided scans.¹² The routine scans relied almost entirely on spotting an endometrioma; the expert scans identified additional disease sites in 97 percent of the patients they detected. A separate study comparing expert operators to less experienced ones found that expert sensitivity for deep infiltrating endometriosis was 77 percent versus 45 percent for first-level operators, and overall accuracy was 88 percent versus 81 percent.¹³

Most of the impressive sensitivity and specificity figures you will find in the research literature come from tertiary centers staffed by gynecological sonographers who perform these scans regularly.¹⁴ If you are sent for a general pelvic ultrasound at a radiology clinic where the technician scans for all sorts of abdominal conditions, the chance of catching non-ovarian endometriosis drops considerably. This does not mean ultrasound is a bad tool. It means that asking specifically for a scan by someone trained in endometriosis assessment, and ideally one following a standardized protocol, makes a material difference in what gets found.

The Role of Bowel Preparation

Something as simple as bowel prep before a transvaginal scan can improve detection of rectosigmoid disease. A prospective study found that after bowel preparation, 22 additional rectosigmoid nodules were identified that had been missed or mistaken for adhesions on an unprepared scan. More than half of the lesions initially classified as adhesions turned out to be actual nodules when the bowel was clearer. The newly detected nodules tended to be smaller, more superficial, or located on the anterior sigmoid wall, areas that bowel gas would have obscured. Patients tolerated the preparation well.¹⁵ If you are scheduled for a dedicated endometriosis ultrasound and are told to prep your bowel beforehand, it is specifically because this step makes the scan more accurate for bowel involvement.

How Ultrasound Compares to MRI

MRI is often perceived as the more advanced imaging option, and for certain locations it has a slight edge. A large meta-analysis found that for deep infiltrating endometriosis in the uterosacral ligaments, the rectovaginal septum, and the bladder, MRI sensitivity tended to be somewhat higher than ultrasound, while specificity was comparable.¹ For rectosigmoid disease, however, the two modalities perform similarly. A head-to-head meta-analysis found pooled sensitivity and specificity for rectosigmoid deep endometriosis of 85 percent and 95 percent for MRI, versus 85 percent and 96 percent for transvaginal ultrasound, essentially a tie.¹⁶

In practice, ultrasound has advantages MRI does not: it is cheaper, more widely available, faster, and allows real-time dynamic assessment like the sliding sign. MRI, meanwhile, offers a wider field of view, which is useful when disease might extend to areas ultrasound cannot easily reach. Many expert centers use both sequentially, starting with a thorough transvaginal ultrasound and adding MRI when the picture is incomplete or surgery is being planned.

Clinical Guidelines Have Shifted Away from Surgery for Diagnosis

For decades, laparoscopy was considered the only definitive way to diagnose endometriosis. You had to go into the operating room, look around, and biopsy tissue. The European Society of Human Reproduction and Embryology (ESHRE) guideline now reflects a significant shift: laparoscopy is recommended only for patients whose imaging is negative and whose empirical treatment has been unsuccessful or inappropriate.¹⁷ This means that in current practice, imaging — and in particular transvaginal ultrasound — is increasingly the first-line diagnostic approach. You are more likely to be sent for an expert ultrasound now than you would have been a decade ago, and less likely to be told that surgery is the only way to know for sure.

This shift has practical implications. It means that getting a high-quality scan early in your diagnostic journey matters more than ever. It also means that if you have symptoms strongly suggestive of endometriosis and your scan is negative, a clinician can still start empirical treatment (typically hormonal therapy) without requiring surgical confirmation first. The scan informs but does not solely determine the diagnosis.

Surgical Mapping and Preoperative Planning

When surgery is planned, a detailed preoperative ultrasound map of where endometriosis has spread becomes critical. The International Deep Endometriosis Analysis (IDEA) consensus laid out a standardized approach to scanning and reporting, covering specific terms and measurements for each pelvic compartment.¹⁸ A structured ultrasound report using this protocol has been shown to correlate well with laparoscopic findings, helping surgeons plan what approach to use and which specialists need to be in the operating room.¹⁹

For rectal endometriosis specifically, combining transvaginal ultrasound with endorectal ultrasound has been shown to improve the accuracy of preoperative assessments, helping determine whether a patient will need bowel shaving, disc excision, or a full segmental resection.²⁰ Endorectal ultrasound can also identify how deeply a nodule has infiltrated the bowel wall layers, and one study found that a thickness cutoff above about 5 millimeters predicted the need for bowel resection with 76 percent sensitivity and 81 percent specificity.²¹ Knowing this before surgery means fewer surprises in the operating room and better outcomes for the patient.

Symptom Information Improves Diagnostic Accuracy

An interesting finding from recent research is that the examiner’s knowledge of your symptoms actually improves how well they read the scan. A study examining uterosacral ligament endometriosis found that when sonographers had access to a patient’s pain scores, they correctly identified disease more often than when scanning without that information. There was also a strong correlation between ligament thickness, nodule identification, and patient-reported pain.²² This means that if you are going for a dedicated endometriosis scan, being specific about your symptoms — where the pain is, when it worsens, whether you have pain with bowel movements or intercourse — genuinely helps the person scanning you know where to look more carefully.

Abdominal Wall Endometriosis Near Surgical Scars

Endometriosis can also implant in the abdominal wall, particularly along cesarean section scars. This is a different situation from pelvic disease, and it calls for a different ultrasound approach. High-frequency surface ultrasound, rather than transvaginal, is used here. The typical appearance is an irregularly shaped, dark nodule sitting within the abdominal wall layers near the scar, often with bright spots from fibrotic tissue and spiky margins. In one case series, about three-quarters of scar endometriomas showed this characteristic pattern.²³ High-frequency color Doppler ultrasound is considered a practical and effective way to diagnose these lesions without needing MRI or biopsy.²⁴

Adolescents and Patients Who Cannot Have Transvaginal Scans

For adolescents who are not sexually active, transvaginal ultrasound may not be appropriate or acceptable. Transabdominal pelvic ultrasound can still pick up larger endometriomas but has limited ability to detect deep disease. Transrectal ultrasound offers a middle ground: it provides a view similar to transvaginal scanning and has been reported to achieve comparable diagnostic accuracy for deep endometriosis, making it a useful alternative in younger patients where transvaginal access is not feasible. If an adolescent has significant pelvic pain, cyclical bowel or bladder symptoms, or other red flags, pursuing one of these alternative routes rather than settling for a transabdominal scan alone is worthwhile.

Artificial Intelligence Is Starting to Close the Expertise Gap

One of the most promising developments in ultrasound for endometriosis is AI-assisted image analysis. A systematic review of AI applications to endometriosis ultrasound found that deep learning models achieved diagnostic accuracies between 89 and 93 percent, with sensitivity ranging from about 78 to 92 percent.²⁵ Perhaps more meaningfully, a recent study showed that when less experienced sonographers were given AI assistance, their diagnostic performance for deep infiltrating endometriosis improved substantially. One junior sonographer’s area under the curve jumped from about 0.75 to 0.88, and their sensitivity rose from 77 percent to 94 percent.²⁶

These are still early-stage tools, and none are in routine clinical use at most hospitals. But they point toward a future where the massive operator-expertise gap described earlier could narrow. If AI can bring a general sonographer’s detection rate closer to that of a tertiary-center specialist, it would change access to accurate endometriosis diagnosis for patients who live far from expert centers.

The Cost Argument for Better Ultrasound Sooner

There is also a financial case for using expert endometriosis ultrasound early in the diagnostic pathway. A cost analysis modeled what happens when dedicated deep-endometriosis transvaginal ultrasound is integrated into the standard workup for women with suspected disease, compared to the conventional pathway that relies on general imaging and often delays diagnosis. The model estimated that for every thousand women investigated, the ultrasound-first pathway saved roughly 925,000 Australian dollars annually in healthcare costs by reducing unnecessary surgeries, repeat consultations, and delayed treatments.²⁷ The long diagnostic delay that many people with endometriosis experience — often years between symptom onset and diagnosis — is not just frustrating. It is expensive for the healthcare system too, and better first-line imaging is one of the more straightforward ways to shorten it.