Removing Colorectal Cancer Tissue Through the Vagina Instead of an Abdominal Cut: What a Review of 823 Patients Found

Table of Contents

• Key Points
• Why This Research Matters
• The Two Surgical Approaches Being Compared
• How the Researchers Conducted the Review
• Which Studies Were Included
• Key Finding 1: Fewer Postoperative Complications
• Key Finding 2: Pathology Results Were Equivalent
• Key Finding 3: No Difference in Cancer Survival
• Key Finding 4: Faster, Less Painful Recovery
• Subgroup and Meta-Regression Analyses
• Publication Bias, Risk of Bias, and Certainty of Evidence
• What This Means for Patients
• Limitations: What This Review Cannot Prove
• Practical Points for Patients Considering This Option
• Frequently Asked Questions
• Source Information

Key Points

• A review of 10 observational studies with 823 patients found transvaginal specimen extraction (TVSE) may reduce postoperative complications compared with an abdominal cut.
• Cancer control appeared equivalent: no significant difference in lymph nodes removed, margins, overall survival, or disease-free survival, though evidence certainty was low.
• All included studies were observational, not randomised, so the authors state conclusions remain inconclusive and call for higher-quality trials.
• TVSE is not suitable for every patient; it is generally not offered to those who are not sexually active or have certain gynaecological conditions, and has no equivalent for men.

Why This Research Matters

Colorectal cancer (cancer of the colon or rectum) is the most common tumour of the digestive system worldwide. The numbers are striking. According to the World Health Organization (WHO), cancer is now the leading cause of death for people under the age of 70 in most countries.

Globally, about 1,880,725 new cases of colorectal cancer were diagnosed in 2020, and 915,880 people died from the disease that year. In China alone, 2022 brought 517,100 newly confirmed cases and 240,000 deaths. Colorectal cancer is now the second most common cancer in China after lung cancer. It ranks fourth for deaths, behind lung, liver, and gastric (stomach) cancer.

Surgery is the cornerstone of treatment that aims to cure colorectal cancer. When it is technically possible, completely removing the primary tumour along with the surrounding lymph nodes (small immune glands that can harbour cancer cells) is central to achieving long-term survival.

But the goals of cancer surgery have widened. As minimally invasive (keyhole) techniques have improved, surgeons no longer focus only on removing the tumour completely. They also want patients to recover quickly and to keep a good cosmetic appearance. Traditional keyhole surgery still needs a small extra abdominal cut to pull the removed tissue out.

That is where natural orifice specimen extraction (NOSE) comes in. NOSE surgery removes the resected tissue through one of the body's natural openings, so no extra abdominal incision is needed. The vagina is one such natural opening, and it is widely used for this purpose in laparoscopic colorectal cancer surgery.

Despite growing interest, big gaps in knowledge remain. Most evidence comes from single-centre studies with differing designs, and clinicians still disagree about how safe the transvaginal route is. This review was designed to fill that gap.

The Two Surgical Approaches Being Compared

Both groups in this analysis had the same main operation: laparoscopic radical surgery for colorectal cancer. The difference lay entirely in how the removed tissue left the body.

• LAP group (transabdominal extraction): the surgical specimen was removed through a small auxiliary incision (an extra cut) in the abdominal wall.
• TVSE group (transvaginal specimen extraction): the surgical specimen was removed through the vaginal route, leaving the abdominal wall uncut.

The researchers compared these two groups on four broad fronts: postoperative complications, pathological results, cancer-related survival, and short-term recovery measures.

How the Researchers Conducted the Review

This was a systematic review and meta-analysis (a study that pools the results of many individual studies into one combined estimate). It followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol, the international standard for this type of work.

The review protocol was registered in advance on PROSPERO, an international prospective register of systematic reviews, under registration number CRD420261341219.

The researchers searched PubMed/MEDLINE, Web of Science, and EMBASE, plus other relevant sources. The search ran from each database's inception up to 16 August 2025, and was limited to English-language publications. Keywords included "natural orifice specimen extraction", "transvaginal specimen extraction", "colorectal cancer", "rectal cancer", and "colon cancer". Unpublished studies and gray literature (research not published in peer-reviewed journals) were excluded.

Two researchers screened the literature independently. Where they disagreed, a third researcher joined the discussion until consensus was reached. The whole process was managed using EndNote reference-management software.

What made a study eligible

1. Patients had colorectal cancer without distant metastasis (spread to other organs) and were eligible for radical surgery.
2. All patients underwent laparoscopic radical surgery, with specimens removed either through a small abdominal incision (LAP) or through the vagina (TVSE).
3. Studies reported on complications, pathology findings (proximal margin, distal margin, lymph node count), cancer outcomes (disease-free survival and overall survival), or short-term outcomes (hospital stay, pain score, additional painkillers, time to passing gas, time to first meal).
4. Study types allowed were randomised controlled trials (RCTs, where patients are assigned by chance) and cohort studies.

Studies were excluded if they were ongoing trials, lacked adequate information or follow-up data, involved other cancer types, or were abstracts, meta-analyses, reviews, comments, or letters to the editor.

Data extraction and quality assessment

Two reviewers independently extracted data using a standardised form. When survival data were not reported directly, the researchers read them off Kaplan-Meier curves (the step-shaped graphs that show survival over time) using software called Engauge Digitizer version 11.3. They then used Tierney's method to calculate the hazard ratio (HR).

Quality was assessed in three ways. The Newcastle-Ottawa Scale (NOS) judged cohort studies, with a score of 7 or higher defined as high quality. The ROBINS-I tool (Risk of Bias in Nonrandomized Studies of Interventions) assessed bias risk. The GRADE framework rated the certainty of the evidence.

Statistical approach

For continuous variables the researchers calculated weighted mean differences (WMDs). For yes/no outcomes they calculated odds ratios (ORs). Heterogeneity — how much the studies disagreed with one another — was measured with Cochran's Q test and the I-squared (I²) statistic. When the Q test P-value was below 0.10 or I² exceeded 50%, they used a random-effects model; otherwise they used a fixed-effects model.

They also ran sensitivity analyses (removing one study at a time to test stability), subgroup analyses, meta-regression, and publication-bias checks using funnel plots plus Begg's and Egger's tests. All statistics were computed in STATA version 12.0, and a two-sided P-value below 0.05 was the threshold for statistical significance.

Which Studies Were Included

The search found 918 articles: 302 from PubMed, 314 from Web of Science, 301 from EMBASE, and 1 from other sources. After removing 457 duplicates, the researchers excluded a further 439 articles based on their titles and abstracts.

Full-text assessment removed 12 more studies, for these reasons:

• Irrelevant outcomes — 1 study
• Abstract or case report — 2 studies
• Ongoing trials — 1 study
• No relevant comparisons — 7 studies
• Used other specimen extraction methods — 1 study

In the end, 10 studies with a total of 823 patients entered the analysis. Their sample sizes ranged from 17 to 140 patients. The studies were published between 2011 and 2025. Six came from China, two from Korea, one from the USA, and one from Russia. Nine were retrospective cohort studies (looking back at existing records) and one was a case-control study.

Individual studies and their quality scores were as follows:

• Gao Gengmei (2019, China, rectal cancer, retrospective cohort): 42 patients — 14 TVSE, 28 LAP; robotic surgery; NOS score 8
• Wei Zheng (2022, China, rectal cancer, retrospective cohort): 100 patients — 52 TVSE, 48 LAP; laparoscopic; NOS 8
• Gengmei Gao (2020, China, sigmoid and rectum, retrospective cohort): 100 patients — 45 TVSE, 45 LAP; laparoscopic; NOS 8
• Xing-Wang Li (2019, China, right colon cancer, retrospective cohort): 62 patients — 31 TVSE, 31 LAP; laparoscopic; NOS 8
• Mingguang Zhang (2022, China, colon cancer, retrospective cohort): 140 patients — 70 TVSE, 70 LAP; laparoscopic; NOS 8
• Hye Jin Kim (2014, Korea, left colorectal cancer, prospective cohort): 116 patients — 58 TVSE, 58 LAP; laparoscopic; NOS 9
• Hongxin Yu (2023, China, right colon cancer, retrospective cohort): 138 patients — 46 TVSE, 92 LAP; robotic; NOS 7
• Ziad T (2014, USA, right colon cancer, retrospective cohort): 40 patients — 20 TVSE, 20 LAP; laparoscopic; NOS 9
• J. S. Park (2011, Korea, right colon cancer, case-control): 68 patients — 34 TVSE, 34 LAP; laparoscopic; NOS 9
• S. K. Efetov (2025, Russia, colon cancer, retrospective cohort): 17 patients — 5 TVSE, 12 LAP; laparoscopic; NOS 7

Key Finding 1: Fewer Postoperative Complications

All 10 studies reported postoperative complications, making this the review's primary outcome. The pooled result favoured the transvaginal approach: odds ratio 0.45 (95% confidence interval 0.30 to 0.67). In plain terms, patients in the TVSE group had roughly 55% lower odds of developing a complication after surgery than patients in the LAP group.

An odds ratio is not the same as an absolute risk, so it does not tell us exactly how many patients per hundred were affected. What it does tell us is the direction and rough size of the difference — and here the difference was clear.

The studies agreed with one another almost perfectly. Heterogeneity was 0.0% (P = 0.935), meaning the studies were consistent, so the researchers used a fixed-effects model.

How serious were the complications?

The researchers used the Clavien-Dindo classification, a standard scale for grading surgical complications. Grade I-II complications are milder (such as a wound infection or a fever needing basic treatment). Grade III-IV complications are serious (such as those needing surgery, intensive care, or organ support).

• Grade I-II complications: no meaningful difference between groups — odds ratio 0.72 (95% CI 0.44 to 1.18). Heterogeneity was low (I² = 15.2%, P = 0.307).
• Grade III-IV complications: no meaningful difference — odds ratio 0.20 (95% CI 0.02 to 1.74). Heterogeneity was 0.0% (P = 0.704). Because the confidence interval crosses 1.0, this result is not statistically significant.

That pattern is important. The overall benefit came mainly from avoiding milder problems rather than from preventing catastrophic ones.

Individual complications

The researchers also pooled each type of complication separately. No heterogeneity appeared, so all of these used a fixed-effects model.

• Intra-abdominal abscess (a pocket of pus inside the belly; 4 studies): odds ratio 0.92 (95% CI 0.23 to 3.76) — no significant difference.
• Anastomotic leakage (a leak at the join where two sections of bowel were reconnected; 5 studies): odds ratio 0.50 (95% CI 0.16 to 1.58) — no significant difference.
• Wound infection (8 studies): odds ratio 0.22 (95% CI 0.09 to 0.55) — a clear reduction in the TVSE group.
• Intestinal obstruction (bowel blockage; 7 studies): odds ratio 0.32 (95% CI 0.13 to 0.77) — a clear reduction in the TVSE group.

The wound-infection finding makes intuitive sense. With no abdominal incision to heal, there is no abdominal wound to become infected.

Key Finding 2: Pathology Results Were Equivalent

A cancer operation is only worthwhile if the cancer is removed properly. The researchers therefore compared the actual tissue removed under the microscope.

Lymph nodes collected. Nine studies reported this. At first, the analysis showed a high level of disagreement between studies (I² = 94%, P = 0.00), so a random-effects model gave a weighted mean difference of 1.87 (95% CI -1.09 to 4.83). The authors caution that with such high heterogeneity, this result should be read carefully.

A sensitivity analysis (removing one study at a time) identified the study by Wei Zheng and colleagues as the main source of the inconsistency, with results clearly diverging from the rest. After removing that single study, heterogeneity dropped to 0.0% (P = 0.937), and the fixed-effects weighted mean difference was 0.54 (95% CI -0.27 to 1.34). The overall conclusion did not change, which supports the robustness of the finding: the number of lymph nodes removed was essentially the same in both groups. This matters because lymph node count is a recognised marker of how thorough a cancer operation was.

Proximal resection margin (the cuff of healthy tissue above the tumour). The two groups were comparable, with no statistically significant difference: weighted mean difference -0.36 (95% CI -1.23 to 0.52) and no heterogeneity between studies.

Distal resection margin (the cuff of healthy tissue below the tumour). The pooled result showed a marginally smaller margin in the TVSE group: weighted mean difference -0.27 (95% CI -0.50 to -0.05), with low heterogeneity. Because the confidence interval only just excludes zero, the authors note this small numerical difference may not carry any meaningful clinical implications.

Key Finding 3: No Difference in Cancer Survival

Cancer safety is the question patients care about most, and here the news was reassuring — the two approaches performed equally well.

Overall survival (OS) — how long patients lived after treatment. Five articles covering 530 patients reported this. Heterogeneity was low (I² = 28.5%, P = 0.231), so a fixed-effects model was used. The hazard ratio was 0.76 (95% CI 0.43 to 1.37). Because this confidence interval includes 1.0, there was no statistically significant difference between the TVSE and LAP groups.

Disease-free survival (DFS) — how long patients lived without the cancer coming back. Five papers covering 546 patients reported this, with no heterogeneity at all (I² = 0.0%, P = 0.994). The fixed-effects hazard ratio was 1.22 (95% CI 0.72 to 2.08). Again, the confidence interval crosses 1.0, so there was no statistically significant difference.

In short, the transvaginal route did not appear to compromise cancer control. It is worth knowing, though, that because the confidence intervals are wide, the analysis cannot rule out a modest benefit or a modest harm in either direction.

Key Finding 4: Faster, Less Painful Recovery

This is where the transvaginal approach showed its most consistent advantages.

• Time to first meal (5 studies): high heterogeneity was present (I² = 70.6%, P = 0.009), so a random-effects model was used. The weighted mean difference was -0.21 (95% CI -0.35 to -0.06), indicating a slightly earlier return to eating. After removing the outlying study by Wei Zheng, heterogeneity fell to 10.2% (P = 0.342) and the fixed-effects result was -0.30 (95% CI -0.41 to -0.20), confirming the finding was robust.
• Hospital stay (9 studies): the pooled result suggested a shorter stay in the TVSE group, with a weighted mean difference of -1.93 days (95% CI -2.81 to -1.06). Heterogeneity was high (I² = 87.8%, P = 0.00), so this should be interpreted with some caution.
• Pain score (8 studies): substantial heterogeneity was present (I² = 81.9%, P = 0.00), so a random-effects model gave a weighted mean difference of -1.29 (95% CI -1.54 to -1.03), meaning pain scores tended to be lower in the TVSE group.
• Additional painkillers (6 studies): no heterogeneity was seen (I² = 0.0%, P = 0.524), so a fixed-effects model gave an odds ratio of 0.47 (95% CI 0.32 to 0.70), indicating patients in the TVSE group were less likely to need extra analgesia.
• Time to first flatus (passing gas, a sign the bowel has woken up; 7 studies): the pooled data showed a shorter time in the TVSE group, weighted mean difference -0.44 (95% CI -0.50 to -0.38), with low heterogeneity (I² = 39.1%, P = 0.131).

Taken together, these measures describe a smoother recovery: less pain, less need for extra pain relief, earlier return of bowel function, earlier eating, and a shorter hospital stay — all without an abdominal wound to heal.

Subgroup and Meta-Regression Analyses

Because pain scores and hospital stay showed high heterogeneity, the researchers dug deeper to find out why. They split the studies into subgroups by tumour location, publication year, sample size, and study region.

The reductions in hospital stay and pain score held up across every subgroup. Residual heterogeneity varied, though:

• For hospital stay, heterogeneity vanished completely in the subgroup with larger sample sizes (I² = 0.0%), suggesting that sample size is an important driver of the disagreement between studies.
• For pain score, heterogeneity fell markedly in four subgroups: left-sided colon studies (I² = 2.6%), studies published before 2020 (I² = 14.0%), non-Chinese studies (I² = 38.0%), and studies with 100 or more patients (I² = 0.0%).

The researchers also ran a meta-regression (a statistical technique that tests whether study results track with particular characteristics) examining patient age, body mass index (BMI), ASA grade I-III (a physical-fitness score used before surgery), and T stage (how deeply the tumour has grown). No significant correlations were found between effect sizes and any of these baseline patient characteristics.

Publication Bias, Risk of Bias, and Certainty of Evidence

The researchers checked whether studies with negative results might have gone unpublished — a common problem called publication bias. The funnel plot for postoperative complications showed slight asymmetry on visual inspection. However, the formal statistical tests were not significant: Begg's test gave p = 0.24 and Egger's test gave p = 0.30. Both are above the 0.05 threshold, so there was no statistical evidence of publication bias.

Using the ROBINS-I tool, the reviewers judged three of the included nonrandomised studies to be at serious risk of bias. This is a meaningful concern, because biased studies can inflate the apparent benefits of a treatment.

Finally, the GRADE framework rated how confident we can be in each result:

• Postoperative complications: moderate certainty
• Clavien-Dindo grade I-II complications: low certainty
• Clavien-Dindo grade III-IV complications: low certainty
• Overall survival: low certainty
• Disease-free survival: low certainty

Only the main complication outcome reached "moderate" certainty. Everything else was rated low, which means further research could change these estimates.

What This Means for Patients

The authors conclude that the TVSE group may be linked to lower postoperative illness and more favourable recovery than the LAP group. No obvious unfavourable effect on cancer safety was seen.

For a patient facing colorectal cancer surgery, this translates into a few practical points worth discussing with a surgical team:

1. Removing the specimen through the vagina avoids an extra abdominal cut, which appears to reduce wound infections and bowel obstruction.
2. Recovery measures — pain, painkiller use, bowel recovery, eating, and length of stay — all trended in favour of the transvaginal route.
3. The cancer operation itself appears just as thorough: similar lymph node counts and similar resection margins.
4. Long-term cancer outcomes appeared equivalent, although the evidence here is of low certainty.

It is also worth understanding the context. Over the past two decades, steady refinement of laparoscopic instruments and surgical techniques has made these approaches possible in the first place. The transvaginal route is one option among several for extracting colorectal specimens, and it is not suitable for every patient — for example, it is generally not offered to patients who are not sexually active or who have certain gynaecological conditions, and it has no equivalent option for men. The original paper does not report suitability criteria in detail.

Limitations: What This Review Cannot Prove

The authors are explicit about the weaknesses of their own work, and patients should weigh these carefully.

• No randomised trials. Every included study was observational. This means patients were not assigned to a surgical approach by chance. Surgeons and patients chose the approach, and that choice may have been influenced by factors such as tumour size, patient anatomy, or overall health — factors that could themselves affect outcomes.
• Small total sample. Only 823 patients across 10 studies, with individual studies ranging from 17 to 140 patients.
• Risk of bias. Three studies were rated as having a serious risk of bias.
• Inconsistent evidence. Four of the five main outcomes were rated as low certainty by the GRADE framework, meaning future research is likely to change the estimates.
• Heterogeneity. Hospital stay and pain score showed substantial disagreement between studies (I² of 87.8% and 81.9% respectively), even though subgroup analyses partly explained it.
• Heavy reliance on a few countries. Six of the 10 studies came from China, which may limit how well the findings apply to other populations.

Because of the observational design of the included studies, the authors state plainly that relevant conclusions remain inconclusive. More high-quality studies are required to validate these results.

Practical Points for Patients Considering This Option

1. Ask about your surgeon's experience with natural orifice specimen extraction. The technique requires specific training, and outcomes improve with volume.
2. Ask whether you are a candidate. Tumour size, location, stage, body habitus, and gynaecological history all influence suitability.
3. Discuss the cancer operation itself, not just the extraction route. What matters most for cure is complete tumour removal with an adequate margin and a thorough lymph node harvest.
4. Expect equivalent cancer outcomes, not better ones. This review found no survival advantage from either approach — the benefit is in recovery, not in cure rates.
5. Understand that the evidence is not yet definitive. Weigh the recovery advantages against the uncertainty, and ask your team what they would recommend for your specific situation.

Frequently Asked Questions

What is transvaginal specimen extraction (TVSE)?

TVSE is a way to remove colorectal cancer tissue through the vagina after laparoscopic surgery, instead of making an extra cut in the belly wall. The main cancer operation is the same; only the route for taking out the specimen differs. This review of 10 studies and 823 patients compared TVSE with removal through a small abdominal incision.

What are the recovery benefits of TVSE compared with an abdominal cut?

Patients in the TVSE group had roughly half the odds of a postoperative complication (odds ratio 0.45), lower pain scores, needed fewer extra painkillers, and left hospital about two days sooner on average. They also passed gas and returned to eating slightly earlier. These findings come from a review of 10 observational studies, so they are not conclusive.

What does an odds ratio of 0.45 for complications mean?

An odds ratio of 0.45 means the TVSE group had about 55% lower odds of developing a complication after surgery compared with the abdominal extraction group. It does not tell you the exact number of patients affected per hundred. The result was statistically significant, but the review authors note that the evidence is not definitive.

Who might be suitable for transvaginal specimen extraction?

The original paper does not report detailed suitability criteria. The review notes that the transvaginal route is not suitable for every patient—for example, it is generally not offered to patients who are not sexually active or who have certain gynaecological conditions, and there is no equivalent option for men. Tumour size, location, stage, body habitus, and gynaecological history all influence suitability.

What are the risks or limitations of the evidence for TVSE?

All 10 included studies were observational, not randomised, so conclusions remain inconclusive. Three studies had a serious risk of bias. Four of five main outcomes were rated low certainty by GRADE. Hospital stay and pain score showed substantial disagreement between studies. Six of 10 studies came from China, which may limit how well findings apply to other populations.

What should I ask my surgeon about this option?

Ask about your surgeon's experience with natural orifice specimen extraction, as outcomes improve with volume. Ask whether you are a candidate based on tumour size, location, stage, body habitus, and gynaecological history. Discuss the cancer operation itself, not just the extraction route. Expect equivalent cancer outcomes, not better ones, and understand the evidence is not yet definitive.

If my surgeon recommends removing colorectal cancer tissue through the vagina instead of an abdominal cut, when should I get a second opinion?

Because every included study was observational rather than randomised, the authors state that conclusions remain inconclusive and call for higher-quality trials. Four of five main outcomes were rated low certainty, and three studies carried serious risk of bias. A second opinion can help you weigh the recovery advantages against that uncertainty, confirm whether you are a candidate, and discuss the cancer operation itself rather than only the extraction route. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Transvaginal specimen extraction in laparoscopic surgery for colorectal cancer: a systematic review and meta-analysis.

Authors: Hao T, Bo Y, Wuyi W.

Publication: BMC Surgery, 2026; volume 26, article 549. DOI: 10.1186/s12893-026-03926-9

Study registration: PROSPERO, registration number CRD420261341219

Publication type: Open Access research article, published under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

This patient-friendly article is based on peer-reviewed research. It reports the findings of a systematic review and meta-analysis of observational studies; it is not medical advice. Patients should discuss their individual treatment options with a qualified surgical and oncology team.