What Does Uterine Cancer Look Like on Ultrasound?

Uterine cancer, most commonly endometrial cancer, typically appears on transvaginal ultrasound as a thickened, irregularly textured endometrial lining with abnormal blood vessel patterns. In postmenopausal women, an endometrial thickness greater than 4 mm on ultrasound is the classic red flag that prompts further investigation, though the picture is rarely that simple. The features that make sonographers suspicious involve more than just thickness, and what looks alarming on the screen does not always turn out to be cancer.

Endometrial Thickness and the 4 mm Threshold

For postmenopausal women with vaginal bleeding, measuring the endometrial stripe on transvaginal ultrasound is the standard first step. An endometrial thickness of 4 mm or less carries a greater than 99 percent negative predictive value for endometrial cancer, meaning that a thin lining reliably rules the disease out in that population.¹ That number is reassuring but not perfect. In one retrospective cohort study of women with confirmed endometrial cancer, about 5.6 percent of cancers presented with a thin endometrium at or below 4 mm, and this false-negative rate was roughly equal for both common and rarer subtypes of the disease.² So a thin lining substantially lowers the odds of cancer, but it does not eliminate them entirely.

In premenopausal women the picture is murkier, because the endometrium changes thickness throughout the menstrual cycle. The well-validated 4 mm cutoff does not apply here. Research on premenopausal women with abnormal uterine bleeding has suggested a cutoff of roughly 9 mm, with a sensitivity around 69 percent and a specificity near 87 percent. That cutoff produced a high negative predictive value of about 98 percent, meaning it was good at ruling cancer out when the lining was thinner, but it missed a larger fraction of cancers than the postmenopausal threshold does.³ This is one reason clinicians rely more heavily on biopsy than on ultrasound alone in younger women with concerning bleeding.

It is also worth noting that an incidentally discovered thick endometrium in a postmenopausal woman who has no bleeding does not automatically call for a workup. Clinical guidelines recommend an individualized assessment based on risk factors rather than reflexive biopsy for every measurement above 4 mm.¹

What the Texture Looks Like

Thickness is only the starting point. Once the endometrium measures above the threshold, the sonographer looks closely at the texture of the lining, which is described by its echogenicity and internal structure. The International Endometrial Tumor Analysis (IETA) group established a standardized vocabulary for describing what the endometrium looks like on ultrasound, covering features like whether the tissue appears uniformly bright or dark, whether it contains cystic spaces, whether its borders are smooth or irregular, and whether the junction between the endometrium and the surrounding muscle wall is disrupted.⁴

Endometrial cancer tends to produce a heterogeneous pattern on ultrasound, meaning the tissue looks uneven or mixed rather than uniform. In an older but influential study of postmenopausal women, a heterogeneous-looking endometrium was found in seven out of eleven cases where such a pattern was seen, making heterogeneity strongly suggestive of carcinoma in that group.⁵ Cancer also tends to distort the normally smooth boundary between the endometrium and the myometrium (the muscular wall of the uterus). When that boundary appears ragged or interrupted, it raises suspicion for invasion into the muscle, which has staging implications.

By contrast, benign conditions often look different. Endometrial polyps frequently contain small cystic spaces that are visible on ultrasound, and the presence of those spaces is a strong predictor of polyps rather than cancer.⁵ Hyperplasia, a precancerous thickening of the lining, more often appears uniformly bright. These distinctions are not foolproof, but they help clinicians weigh how urgently biopsy is needed.

A large prospective multicenter study applied the IETA terminology to over 1,700 women with biopsy-confirmed endometrial cancer, systematically describing how the cancer looked in relation to its stage, grade, and cell type.⁶ That kind of standardized classification has been important for teaching sonographers what patterns to flag. The IETA system helps ensure two different clinicians describe the same image in the same way, reducing the subjectivity that has historically been a weak point of ultrasound.⁷

Blood Flow Patterns on Doppler

Color Doppler ultrasound adds another layer to the picture by showing blood flow within and around the endometrium. Cancerous tissue tends to recruit new blood vessels to feed its growth, and those vessels behave differently from normal ones. On Doppler, endometrial cancers often show multiple vessels originating from a single area or from scattered points throughout the thickened lining. A study of endometrial cancer patients found that the presence of multiple global vessels and a moderate or high color score were both significantly more common in higher-stage tumors, higher-grade tumors, and larger tumors.⁸ In practical terms, more blood flow on Doppler suggests a more advanced or aggressive cancer.

Doppler can also help distinguish malignant from benign thickening. In a study comparing postmenopausal women with benign versus malignant endometrial polyps, the malignant group showed much richer blood flow. About 70 percent of malignant polyps had moderate to high-grade blood flow, compared with roughly 29 percent of benign ones. The blood vessels in malignant lesions also had lower resistance, meaning the blood flowed more freely through them.⁹ These differences are useful clues, though none is specific enough to replace biopsy.

Assessing How Far Cancer Has Spread

Once endometrial cancer is confirmed by biopsy, a key question is how deeply it has grown into the uterine wall. Tumors confined to the inner half of the myometrium are staged differently from those that have pushed into the outer half, and staging directly affects surgical decisions and prognosis. Ultrasound plays a significant role here. On the image, the examiner looks at the ratio of the tumor’s thickness to the total thickness of the uterus at the point of deepest invasion. If the boundary between the endometrium and the myometrium is disrupted past the halfway mark, deep myometrial invasion is suspected.

A systematic review and meta-analysis comparing three-dimensional transvaginal ultrasound to MRI for detecting deep myometrial invasion found that the two methods performed similarly. Pooled sensitivity was around 77 percent for 3D ultrasound and 80 percent for MRI, and statistical analysis showed no meaningful difference between them.¹⁰ A separate systematic review reached a similar conclusion, reporting pooled sensitivity of about 80 percent for 3D ultrasound and 81 percent for MRI, with the biggest gain coming when the two methods were combined.¹¹ Because ultrasound is cheaper, faster, and more accessible than MRI, these findings support its growing role in preoperative staging.

Assessing whether cancer has invaded the cervix is another important staging question. This turns out to be harder on ultrasound, and the examiner’s expertise matters a lot. In a study comparing ultrasound experts to general gynecologists, the experts achieved meaningfully higher sensitivity and specificity for detecting cervical stromal invasion, and the agreement between different experts was much better than the agreement between different gynecologists.¹² Intrauterine sonography, a technique in which a small probe is placed directly inside the uterine cavity, has shown promising specificity for cervical invasion in small studies, though it remains less widely used.¹³

When Ultrasound Misses or Misleads

Ultrasound has genuine blind spots. Several factors can degrade image quality or mimic cancer where none exists.

Uterine fibroids are one of the most common confounders. Large fibroids can distort the uterine cavity, making it difficult to measure the endometrial thickness accurately or to visualize the lining at all. The endometrium may appear thicker than it truly is, or focal abnormalities within it may be hidden behind fibroid shadows. One study specifically examined the impact of fibroids and non-endometrioid cancer subtypes on ultrasound performance and found that both could degrade the test’s accuracy.¹⁴

The position and orientation of the uterus also matter. Most women have an anteverted uterus (tilted forward), which is easy for a transvaginal probe to image. But a retroverted (tilted backward) or axially positioned uterus can be harder to visualize clearly, and this is compounded by obesity. A narrative review of the topic found that transabdominal ultrasound in particular showed lower diagnostic reliability in retroverted or axial uteri and in patients with obesity or focal lesions.¹⁵ Other factors that can compromise accuracy include very small invasions of just a millimeter or two (too subtle for imaging to detect), the presence of adenomyosis, recent intrauterine procedures, and active uterine bleeding at the time of the scan.¹⁶

Adenomyosis, a condition in which endometrial tissue grows into the muscular wall, poses a particular diagnostic trap. In rare cases, cancer can arise directly within adenomyosis, making it almost impossible to detect on conventional imaging before surgery. A case report documented a 34-year-old woman whose cancer arose within adenomyosis and was only confirmed after surgical removal because preoperative imaging could not distinguish the malignancy from the surrounding benign adenomyosis.¹⁷

The Sarcoma Problem

Endometrial cancer is the most common type of uterine cancer, but it is not the only one. Uterine sarcomas, which arise from the muscle or connective tissue of the uterine wall rather than the lining, present a much harder challenge on ultrasound. The most feared of these is leiomyosarcoma, a malignant tumor that can look nearly identical to a benign fibroid on imaging. Fibroids are extremely common, and the vast majority are harmless, so there is no routine imaging protocol designed to screen for the rare sarcoma hiding among them. Imaging differentiation between the two is frequently challenging because their ultrasound features overlap considerably, and a misdiagnosis can delay treatment with serious consequences.¹⁸ Some suspicious features, like rapid growth of a known fibroid, heterogeneous internal echoes, and central necrosis, can raise suspicion, but none is specific enough to be reliable. MRI is typically added when sarcoma is a concern, though even MRI has limitations.

Tamoxifen and the Misleading Thick Endometrium

Women taking tamoxifen, a drug commonly used in breast cancer treatment, deserve a separate mention. Tamoxifen has estrogen-like effects on the uterus and can cause the endometrium to thicken substantially, even in the absence of disease. This creates a diagnostic dilemma: the endometrium may look worryingly thick on ultrasound, but biopsies often turn up benign findings. In one study of tamoxifen-treated women with endometrial thickness of 8 mm or more, the most common biopsy results were normal endometrium and polyps, with only two cancer cases among 36 biopsied patients.¹⁹

More recent research has tried to refine the threshold at which tamoxifen users should undergo further investigation. A study examining breast cancer patients on tamoxifen found that endometrial thickness was significantly greater in those with underlying pathology than in those without, in both premenopausal and postmenopausal groups. The proposed cutoff values differed from the standard 4 mm used in non-tamoxifen populations: roughly 9.5 mm for premenopausal patients and 5.5 mm for postmenopausal patients.²⁰ These adjusted thresholds reflect the drug’s confounding effect and help avoid unnecessary biopsies while still catching real problems.

Newer Ultrasound Technologies

Standard grayscale and color Doppler ultrasound remain the workhorses, but several newer techniques are being explored to improve accuracy.

Contrast-enhanced ultrasound (CEUS) involves injecting microbubble contrast into a vein, which lights up the blood vessels inside the uterus in real time. Cancerous tissue typically enhances earlier and more intensely than benign tissue, and 3D contrast imaging can reveal tortuous, irregular blood vessels clustered in groups. One systematic review reported an average accuracy of 81 percent for CEUS in staging endometrial carcinoma.²¹ A more recent study focused specifically on early vascular architecture seen during the first moments after contrast injection, finding that the presence of a “vascular supply” pattern had a sensitivity of about 96 percent and accuracy of roughly 84 percent for identifying malignant endometrial lesions.²²

Microvascular flow imaging (MV-Flow) is another emerging tool that detects very small, low-velocity blood vessels that standard color Doppler misses. In one study, MV-Flow detected blood flow in all malignant lesions, compared with about 86 percent detection using conventional Doppler. The technique also produced a higher diagnostic accuracy, with an area under the curve of 0.87 for its quantitative parameter when used to differentiate benign from malignant endometrial lesions.²³ These technologies are not yet standard in every clinic, but they represent the direction the field is heading.

Artificial Intelligence in Ultrasound Interpretation

One of the persistent weaknesses of ultrasound is that its accuracy depends heavily on who is holding the probe and reading the image. AI-based tools are being developed to reduce this variability. Machine learning models trained on ultrasound images can extract features from the image that the human eye might miss, including subtle textural and spatial patterns embedded in the pixel data.

One recent model combined traditional image feature analysis with deep learning from convolutional neural networks, aiming to improve the ability to distinguish cancerous endometria from benign thickening in postmenopausal women. The researchers reported that the hybrid approach improved specificity and reduced false positives compared with individual methods alone.²⁴ Another study combined image-derived features with clinical data like patient age, bleeding history, and menopausal status to build a triage model for identifying high-risk endometrial lesions.²⁵ Models combining these features with the IETA-based descriptors have also been explored for telling polyps with hyperplasia apart from actual cancer on ultrasound.²⁶

These tools are still in the research and validation phase, not yet part of routine clinical workflows. But given how operator-dependent conventional ultrasound is, the prospect of an AI assistant flagging suspicious features in real time could eventually change how the test performs in average practice settings rather than just in the hands of experts.

Why Ultrasound Is Not Used for Screening

Given that ultrasound can detect endometrial cancer, you might wonder why it is not offered as a routine screening test the way mammography is for breast cancer. The answer comes down to the balance between benefit and harm in a low-risk population. Evidence-based assessments have concluded that vaginal ultrasound screening in asymptomatic women is neither medically useful nor cost-efficient for the early detection of endometrial cancer.²⁷ The false-positive rate would be high, leading to unnecessary biopsies, anxiety, and medical costs in women who turned out to be fine. Endometrial cancer fortunately tends to announce itself early through abnormal bleeding, which sends women to their doctor and into the imaging pathway before the cancer has spread far. That symptom-driven approach remains far more effective than screening a population of women who feel well.

The exception involves women at very high risk, such as those with Lynch syndrome, a genetic condition that dramatically raises the lifetime risk of endometrial and other cancers. Even in that group, guidelines vary by institution, and the evidence supporting routine ultrasound surveillance is debated. Saline infusion sonohysterography, a technique in which fluid is instilled into the uterine cavity to improve visualization, has been considered as an enhanced approach, though a small study found that malignant cells were present in the spilled fluid in one out of fourteen cases, raising theoretical concerns about cell dissemination.²⁸ That finding has not been shown to worsen outcomes, but it illustrates how even adjunctive imaging methods in this setting involve tradeoffs that have not been fully resolved.