Accuracy of transvaginal duplex ultrasound compared to catheter venography for the identification of pelvic vein incompetence

Abstract

Objective

Chronic pelvic pain (CPP) is a debilitating condition affecting a quarter of premenopausal women and has been associated with pelvic vein incompetence (PVI). Catheter venography (CV) is the standard investigation, although transvaginal duplex (TVDU) ultrasound has emerged as a promising alternative. The accuracy of TVDU compared to CV is undefined and the optimal reflux time measured by TVDU to diagnose PVI is unknown. This study aimed to establish the diagnostic accuracy of TVDU in those with suspected PVI.

Methods

This cohort study included women enrolled into a randomised controlled trial comparing pelvic vein embolisation to no treatment, who underwent both TVDU and CV. Three consultant vascular interventional radiologists, blinded to TVDU results, independently reviewed catheter venograms. Reflux in bilateral ovarian and internal iliac veins was reported as a binary outcome. Pelvic reflux times, in both supine and semi-standing positions, were retrieved from TVDU reports. Diagnostic accuracy of TVDU was assessed using CV as the comparator. Sensitivity, specificity and area under the ROC curve (AUC) were calculated.

Results

124 pelvic veins were analysed in 31 women. For a diagnostic reflux time threshold of 500 ms, TVDU had a sensitivity of 64% and specificity of 78% (AUI 0.71,95%CI 0.66-0.76). For a threshold of 700 ms, TVDU had a sensitivity of 74% and a specificity of 78% (AUI 0.76,95%CI 0.71-0.80). For a threshold of 700 ms, TVDU performed semi-standing had a sensitivity of 78% and specificity of 78% (AUC of 0.78,95%CI 0.72-0.84) as compared to TVDU performed supine which had a sensitivity of 70% and specificity of 78% (AUC 0.74,95%CI 0.67-0.80).

Conclusions

TVDU demonstrates good diagnostic accuracy for PVI using a 700 ms reflux time threshold. Performing the test in a semi-standing position optimises accuracy. TVDU has potential as a diagnostic tool and may eliminate unnecessary invasive investigations, thus preserving catheter venography for intention to treat.

Keywords

Venous reflux pelvic vein incompetence transvaginal ultrasound duplex ultrasound catheter venography

Introduction

Chronic pelvic pain (CPP) represents a substantial health concern for women, affecting an estimated 24% of the premenopausal population.¹ It is defined by the Royal College of Obstetricians and Gynaecologists (RCOG, U.K.) as non-cyclic persistent, or recurrent lower abdominal or pelvic pain of more than 6 months duration.² CPP substantially impairs quality of life and accounts for 20–30% of gynaecology outpatient appointments within the United Kingdom (U.K.), contributing to a considerable economic impact estimated at €3.8 billion annually across Europe.^3,4 It accounts for 40% of laparoscopies and 12% of hysterectomies in the USA annually.⁵

A key challenge in the management of CPP lies in its diagnosis. Despite comprehensive investigations, often including laparoscopy, up to 55% of women do not receive a definitive diagnosis.⁵ CPP is a symptom complex rather than a distinct disease and often arises secondary to diverse pathologies. Pelvic vein incompetence (PVI) is a proposed aetiological factor in some women with CPP.⁶ Characterised by retrograde flow within the ovarian and/or internal iliac veins, PVI is poorly understood and can be undetected during laparoscopy. Persistent retrograde flow can result in chronic venous hypertension and pelvic varices.^7,8 Neither the RCOG nor the U.K. National Institute of Clinical Excellence (NICE) provide specific PVI management guidance, contrasting with testicular varicocele management within the UK National Health Service (NHS).⁹ In the UK, a lack of robust data is often cited by integrated care boards who withhold funding for women seeking to receive treatment for PVI in the NHS.

Catheter venography (CV) has traditionally served as the diagnostic gold standard for PVI.¹⁰ However, CV is an invasive procedure requiring percutaneous access, typically via the jugular vein. Potential risks include, but are not limited to, bleeding complications, anaphylaxis to the contrast agent, deep vein thrombosis, and post-embolisation syndrome.¹¹ Its interpretation is guided by accepted practice but lacks a universally standardised diagnostic criteria and often reliant on radiologist expertise. There is also a resource allocation cost with utilisation of an interventional radiology suite, endovascular consumables and specialists’ time. Transvaginal duplex ultrasound (TVDU) offers a less invasive modality for assessing pelvic venous anatomy and haemodynamic characteristics, including the detection and quantification of venous reflux.¹² Whilst TVDU requires a specialist operator, it is likely to be more time and cost effective than venography and allows release of interventional radiology suite for therapeutic, as opposed to diagnostic, purposes. Like CV, there is no experimental data underpinning its use, although it has been the subject of a Delphi consensus exercise.¹³ Ultimately, robust data regarding the sensitivity and specificity of TVDU for diagnosing PVI are needed.

In addition, there is a need to determine which pathophysiological features identified on TVDU have clinical relevance. Previous studies have nominated greater venous diameter,¹⁴ prolonged reflux time,¹⁵ and the presence of varices as confirmatory features,⁹ heterogeneity that betrays a lack of consensus on robust diagnostic criteria. If TVDU can demonstrate acceptably high sensitivity and specificity versus CV in the diagnosis of PVI, then the clinical risk and unit cost profile of investigating women with CPP may be mitigated. Defining a clinically relevant TVDU-detected reflux time would enhance the modality’s credibility as a useful investigation in women with CPP and suspected PVI.

In that context, the primary objective of this study was to define the sensitivity and specificity of TVDU in against CV, to diagnose PVI in a cohort of pre-menopausal women with CPP. Secondarily, the analysis was repeated using different diagnostic thresholds (500 ms and 700 ms reflux time) to determine their effect on sensitivity and specificity.

Methods

This study was a retrospective analysis of data from a randomised controlled trial (RCT) conducted by our group across two institutions between 2016 and 2020.¹⁶ In this RCT, pre-menopausal women aged 18–54 years with CPP underwent TVDU to determine the presence of PVI. TVDU was considered to show PVI if sustained reflux >500 ms, in any one of the left or right ovarian or internal iliac veins, was identified during Valsalva’s manoeuvre. This was in accordance with our group’s Delphi consensus of British Society of Interventional Radiology members.¹³ Participants meeting this criterion proceeded to CV.

For TVDU, the transvaginal duplex probe was introduced in a supine position. The duration of reflux in bilateral ovarian and internal iliac veins was assessed during Valsalva’s manoeuvre. This was repeated in the semi-standing position, with the patient sat on the edge of the examination couch. Semi-standing positioning mimics the reverse Trendelenburg position, without the need for a specific examination couch, and was utilised with the intention of maximising venous congestion in the pelvis during scanning. TVDU was performed according to a protocol that has previously been described by our group.¹² For each scan, four veins were isolated in the coronal oblique plane: the left ovarian vein (OV), right OV, left internal iliac vein (IV) and the right IIV. The IIV was identified along the lateral boundary of the ovary on each side, extending towards its junction with the external and common iliac veins. The main trunk of the ovarian vein is commonly inaccessible to transvaginal insolation; as such, assessment of peri-ovarian veins was performed to aid interpretation. To visualise the IIVs and OVs, the transducer was angled into the right and left lateral fornices of the cervix. In these studies, the TVDU was not performed to determine any additional intra-abdominal or pelvic pathology, including left renal vein compression.

In all cases, transjugular venography was performed in an interventional radiology suite under local anaesthesia. Using ultrasound guidance, the internal jugular vein was accessed using the Seldinger technique. Selective catheterisation of bilateral ovarian and internal iliac veins was performed, and the presence of venous incompetence was assessed during Valsalva’s manoeuvre. Evidence of reflux, venous dilatation, pelvic varices, and venous plexus communication across the midline or between pelvis and thigh and/or vulvovaginal varices were recorded. Due to the difficulty of standardising catheter tip positioning and deployed force during contrast injection within the ovarian vein, rapidity of contrast flow and reflux time were not measured.

For this analysis, the primary outcome measure was the sensitivity and specificity of TVDU for PVI, using CV as the reference standard. TVDU velocities were compared with the binary PVI diagnosis from CV, for each individual ovarian and internal iliac vein. Secondary outcomes included comparing a TVDU reflux time threshold of 700 ms with a default 500 ms threshold for accuracy against catheter venography. Thirdly, TVDU assessments in supine and semi-standing positions were compared.

TVDU velocity data were retrospectively collected from imaging reports. Anonymised CV images were analysed using Picture Archiving and Communication System (PACS; Sectra Medical, Linkoping, Sweden) by three Consultant Interventional Radiologists who were blinded to TVDU results. They provided a binary overall PVI impression, based on venographic assessment of retrograde flow, vessel diameter, vessel tortuosity and congestion.

Only participants randomised to the no-treatment arm of the original RCT were included in this blinded retrospective analysis. Venographic images from participants who underwent embolisation were excluded to avoid bias in the blinded image assessment from the presence of embolisation coils.

Statistical analysis, including area under the receiver operating characteristic curve (AUC), sensitivity, and specificity calculations, was performed using Prism version 10 (GraphPad Software, Boston, USA); p < .05 was considered as statistically significant.

Results

A total of 124 pelvic veins in 31 women were included in this analysis. Demographics of these women who were randomised to no treatment in the original RCT have been detailed in a prior publication¹⁶ and are summarised in Table 1.

Table 1.

Demographics of 31 women included in the study.

Age (years)	Mean +/− SD	40.1	8.3
Body Mass Index (kg/m²)	Median (IQR)	23.3	22.1-25.9
Gravida	Median (IQR)	2.5	2-3
Parity	Median (IQR)	2	0-4
Smoking status:
Never smoked	%	50
Ex-smoker	%	16.7
Current smoker	%	33.3

SD = standard deviation, IQR = interquartile range.

Using a 500 ms diagnostic reflux threshold, TVDU demonstrated 64% sensitivity and 78% specificity (AUC 0.71, 95% CI 0.66–0.76). In the supine position, TVDU had moderate accuracy for PVI identification (AUC 0.69, 95% CI 0.62–0.76, p < .0001; Figure 1(a)), with 60.83% sensitivity and 77.68% specificity. Accuracy improved in the semi-standing position (AUC 0.73, 95% CI 0.66–0.79, p < .0001; Figure 1(b)), with 77.68% sensitivity and specificity.

Figure 1.

Area Under the Curve plots for 500 ms TVDU diagnostic threshold in (a) Supine and (b) Semi-Standing positions. Caption for Figure 1: AUC = Area under the receiver operating characteristic curve, CI = confidence interval.

The accuracy of TVDU with a 500 ms threshold varied depending on the specific vessel being assessed (Table 2). The right ovarian vein (ROV) showed the lowest accuracy (supine: AUC 0.64, 95% CI 0.49–0.79, p = .076; semi-standing: AUC 0.69, 95% CI 0.55–0.83, p = .016), with 60% sensitivity and 68% specificity in both positions. Notably, the left ovarian vein (LOV) demonstrated relatively high specificity of 90% in both supine and semi-standing positions.

Table 2.

Diagnostic accuracy of TVDU (500 ms threshold) as compared to CV, for specific pelvic veins.

Vein	Position	AUC	95% CI	Sensitivity (%)	Specificity (%)
LIIV	Semi-standing	0.71	0.5681 to 0.8431	63.33	77.78
LIIV	Supine	0.69	0.5493 to 0.8285	60.00	77.78
LOV	Semi-standing	0.78	0.6536 to 0.9016	65.52	90.00
LOV	Supine	0.67	0.5344 to 0.8138	44.83	90.00
RIIV	Semi-standing	0.72	0.5906 to 0.8524	70.97	73.33
RIIV	Supine	0.71	0.6615 to 0.7569	77.42	73.33
ROV	Semi-standing	0.69	0.5469 to 0.8331	70.00	68.00
ROV	Supine	0.64	0.4919 to 0.7881	60.00	68.00

AUC = Area under the receiver operating characteristic curve, CI = confidence interval, LIIV = left internal iliac vein, LOV = left ovarian vein, RIIV = right internal iliac vein, ROV = right ovarian vein.

Increasing the TVDU reflux time threshold to 700 ms improved overall accuracy (74% sensitivity, 78% specificity; AUC 0.76, 95% CI 0.71–0.80). In the supine position, the AUC was 0.74 (95% CI 0.67–0.81, p < .0001), with 70.09% sensitivity and 77.68% specificity (Figure 2(a)). In the semi-standing position, the AUC was 0.78 (95% CI 0.72–0.84, p < .0001), with 78.07% sensitivity and 77.68% specificity (Figure 2(b)).

Figure 2.

Area Under the Curve plots for 700 ms TVDU diagnostic threshold in (a) Supine and (b) Semi-Standing positions. Caption for Figure 2: AUC = Area under the receiver operating characteristic curve, CI = confidence interval.

With the 700 ms threshold, TVDU accuracy also varied by the specific vessel being assessed (Table 3). The ROV again had the lowest accuracy (supine: AUC 0.64, 95% CI 0.54–0.83, p = .020; semi-standing: AUC 0.73, 95% CI 0.59–0.87, p = .004). Of note, the LOV again demonstrated relatively higher specificity of 90% in both supine and semi-standing positions.

Table 3.

Diagnostic accuracy of TVDU (700 ms threshold) as compared to CV, for specific pelvic veins.

Vein	Position	AUC	95% CI	Sensitivity (%)	Specificity (%)
LIIV	Semi-standing	0.77	0.6453 to 0.8992	76.67	77.78
LIIV	Supine	0.76	0.6256 to 0.8855	73.33	77.78
LOV	Semi-standing	0.83	0.7197 to 0.9495	76.92	90.00
LOV	Supine	0.72	0.5820 to 0.8537	53.57	90.00
RIIV	Semi-standing	0.77	0.6421 to 0.8913	80.00	73.33
RIIV	Supine	0.78	0.6620 to 0.9047	83.33	73.33
ROV	Semi-standing	0.73	0.5933 to 0.8724	78.57	68.00
ROV	Supine	0.68	0.5400 to 0.8297	68.97	68.00

Discussion

PVI is a common condition in women, often associated with CPP, and accurate diagnosis is crucial for effective management. There are a significant number of women who endure not only the symptoms of CPP, but also a multitude of invasive investigations and treatments such as diagnostic laparoscopy, hysteroscopy, and hysterectomy. Accordingly, any investigative tool that seeks to broaden the diagnostic spectrum for women with CPP should aim to be low cost and confer minimal risk. Catheter venography (CV) has long been considered the gold standard for diagnosing PVI, but it is an invasive procedure with a greater complication risk and unit cost than TVDU. This study evaluated the diagnostic accuracy of TVDU compared with CV for the identification of PVI in women presenting with chronic pelvic pain.

The primary finding was that TVDU demonstrated good diagnostic accuracy for PVI, with optimal performance observed using a 700 ms reflux time threshold. Furthermore, semi-standing patient positioning yielded superior diagnostic accuracy compared with the supine position. Notably, out of all four possible incompetent veins (left and right for both ovarian and internal iliac veins), assessment of the left ovarian vein specifically demonstrated excellent specificity, potentially reflecting easier acquisition of high-quality anatomical images of the left adnexa. This corroborates a recent finding from a cohort study in China, that demonstrated LOV incompetence on transabdominal ultrasound scan as being an independent risk factor for CPP.¹⁷ Our observation suggests a potential strength of TVDU as a screening modality to guide subsequent definitive investigation and treatment with CV.

TVDU offers a less invasive alternative, but its diagnostic accuracy relative to CV has been subject to debate. A Delphi consensus study highlighted a lack of expert consensus regarding the clinical utility of TVDU, which was attributed to the paucity of robust evidence in the literature.¹³ Our findings address this evidence gap and suggest that TVDU can be a valuable tool in the diagnostic workup of PVI, especially when performed in the semi-standing position with a 700 ms reflux time threshold.

Prior studies investigating TVDU for PVI diagnosis are frequently limited by methodological inconsistencies and small sample sizes, hindering definitive conclusions regarding its diagnostic accuracy.¹⁰ While catheter venography (CV) remains the current gold standard, its interpretation lacks a universally standardised diagnostic protocol. Consequently, the role of TVDU in PVI diagnosis requires further investigation.

Several studies assessing TVDU accuracy are hindered by incomplete venographic verification, precluding accurate specificity calculation. For instance, Halligan et al.¹⁸ used scoring systems for both venography and ultrasound but only included patients with a venographic score above a certain threshold, leading to a sensitivity calculation of only 51% for ultrasound. Similarly, several studies lack clear definition of ultrasound findings, or combine positive and negative agreement, limiting interpretation of their results. Bachar et al.¹⁵ and Creton et al.,¹⁹ focused on the effect of embolisation and provided limited descriptions of their diagnostic work-up. Bachar et al. only performed transabdominal ultrasound (TAUS) and venography on four patients, making any meaningful analysis of diagnostic accuracy impossible.¹⁵ Creton et al. while including 24 patients, focused solely on the anatomical distribution of varices, neglecting detailed analysis of ultrasound findings related to PVI.¹⁹

Park et al.²⁰ assessed 139 patients with suspected PVI who underwent both transabdominal and transvaginal ultrasound. 74 patients were described as having pelvic varices without other pathology, but only 32 of these underwent venography, preventing positive predictive value calculation. In contrast, Barros et al.²¹ conducted a more comprehensive evaluation of TVDU and identified pelvic varices in 150 of 249 participants, compared to 156 by venography. They reported high accuracy with TVDU (96% sensitivity 95% CI 92–99%, 100% specificity lower 99% CI 97%), however these findings contrast with the limitations and inconsistencies observed in other studies.

A previous study by our group set the scene for our current work by demonstrating an association between TVDU, CV and PVI.¹⁶ One other study has reported excellent accuracy for TVDU but utilised a different methodology, involving the comparison of post-embolisation TVDU to pre-treatment TVDU and CV reflux at the time of treatment.²² Ultimately, robust evidence supporting the diagnostic accuracy of TVDU in the literature is scarce.

The observed improvement in diagnostic accuracy with a 700 ms threshold and semi-standing positioning is consistent with the increasing recognition of the importance of patient positioning and haemodynamic factors in the assessment of PVI.¹³ This finding also addresses a significant limitation of CV as a first-line investigation, namely the requirement for patient decubitus positioning. The higher sensitivity observed in the semi-standing position is likely to reflect increased venous pressure and distension, which enhances the visualisation of reflux.

Increasing the reflux time may increase accuracy for several reasons; it will likely better distinguish between pathological and normal/physiological reflux, a longer reflux time may correlate with severity of valve incompetence, and it will reduce the risk of a false negative result.

The diagnostic thresholds studied in the pelvis differ slightly to those used in the lower limb, classically quoted as 500 ms for superficial veins and 1000 ms for deep veins. This may be explained by the fact that pelvic veins differ significantly from deep and superficial veins in the lower limbs in terms of their anatomy, physiology, and hemodynamic characteristics. They are typically thin walled, lack the support of surrounding muscle fascia to the same extent, and are subject to hormonal influences and changes during the menstrual cycle and pregnancy to a greater extent. Interrogation of further potential threshold values beyond 500 ms and 700 ms could be explored further. Our impression is that higher threshold values will likely correlate with severity of pelvic congestion symptoms, but larger studies would be required to investigate this.

A key strength of this study is its utilisation of data derived from a rigorously conducted randomised controlled trial (RCT) in which all participants underwent both TVDU and CV, thereby minimising selection bias. The blinded assessment of venograms further enhances the validity of the study findings. To our knowledge, this is the first study to directly compare the diagnostic accuracy of TVDU and CV for PVI diagnosis using a standardised protocol and blinded assessment.

Limitations of this study include the potential for subjectivity in the interpretation of venograms, which may introduce variability in the diagnosis of PVI. Furthermore, as CV was performed in the original RCT only in participants with positive TVDU findings, the negative predictive value and specificity of TVDU could not be assessed with complete accuracy. This potential introduction of positive verification bias is a key limitation of our work that could be overcome with further validation studies in the future. Pelvic vein diameter is a metric of interest for the diagnosis of PVI, however was not within the original design scope of this analysis. Diameter is an anatomical variable across a normal distribution and therefore potentially only a surrogate marker of pathology. Reflux at >700 ms is a pathological marker of dysfunction independent of confounding factors like height and weight. Correlation between vein diameter on TVDU and reflux on CV could be the subject of further work. Finally, all TVDU scans were performed by a single operator, which may limit the generalisability of the study findings to other settings.

The findings of this study suggest that TVDU performed in the semi-standing position with a 700 ms reflux time threshold may optimise its diagnostic accuracy for PVI. We propose that TVDU should be the primary diagnostic/screening investigation for PVI, with venography withheld until there is an intention to treat. We anticipate that this strategy will reduce the cost of investigating possible PVI, will minimise resource allocation, will avoid exposure to the risks of CV, and will enhance patient experience.

A health economic analysis of this strategy arising from our clinical trial found that this approach is likely to be cost-effective.²³ However, we accept that relative inexperience amongst vascular scientists may limit access to TVDU in the short term. In mitigation, training is not complex, and techniques are in accordance with the principles of lower limb venous imaging which is a core skill. Accessibility of any imaging technique is likely to improve in the context of robust data confirming its value.

Future multi-centre prospective studies involving a diverse range of operators are required to validate these findings and assess the generalisability and operator dependency of TVDU. Based on these findings the implementation of standardised TVDU protocols has the potential to improve the diagnosis and management of PVI in both clinical practice and research settings.

Conclusion

TVDU in the semi-standing position using a 700 ms reflux time threshold demonstrates acceptable diagnostic accuracy for PVI in women with CPP. The deployment of TVDU for investigation, and CV for treatment, may be a more accessible and acceptable management pathway for women with CPP and suspected PVI. Further research is warranted to validate these findings and to refine TVDU protocols to optimise ‘real world’ and clinical trial strategies.

Footnotes

Author contributors

JG and DR principally conceived the idea for the project (substantial contribution).

SB, DS and AG collected and interpreted the imaging data (substantial contribution).

DB and SP collected and analysed the data (substantial contribution).

KN performed all TVDU measurements (substantial contribution).

All authors (DB, SP, DR, KN, SB, DS, AG, JG) contributed to reviewing the manuscript critically, approving the version to be submitted, and agree to be accountable for all aspects of the work (as per ICMJE basis for authorship).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Guarantor

Dr Jonathan Ghosh (Senior Author).

Ethical approval

This was a retrospective analysis of data from a randomised controlled trial that had previously been completed and published with full ethical approval - NRES Committee North West - Greater Manchester East, 17/06/2015, ref 15/NW/0360.

ORCID iD

Daxina Bhatt

Data Availability Statement

Data can be made available upon request.*

References

Latthe

Say

, et al. WHO systematic review of prevalence of chronic pelvic pain: a neglected reproductive health morbidity. BMC Public Health 2006; 6: 177.

RCoOa

Gynaecologists

. The initial management of chronic pelvic pain: green Top Guideline No. 41. Available from: https://www.rcog.org.uk/globalassets/documents/guidelines/gtg_41.pdf (2012).

Howard

. The role of laparoscopy in chronic pelvic pain: promise and pitfalls. Obstet Gynecol Surv 1993; 48(6): 357–387.

Riding

Hansrani

McCollum

. Pelvic vein incompetence: clinical perspectives. Vasc Health Risk Manag 2017; 13: 439–447.

Lamvu

Carrillo

Ouyang

, et al. Chronic pelvic pain in women: a review. JAMA 2021; 325(23): 2381–2391.

Hahn

Steege

. Pelvic congestion syndrome-associated pelvic pain: a systematic review of diagnosis and management. Obstet Gynecol Surv 2010; 65(5): 332–340.

Phillips

Deipolyi

Hesketh

, et al. Pelvic congestion syndrome: etiology of pain, diagnosis, and clinical management. J Vasc Interv Radiol 2014; 25(5): 725–733.

Meissner

Khilnani

Labropoulos

, et al. The symptoms-varices-pathophysiology classification of pelvic venous disorders: a report of the American vein & lymphatic society international working group on pelvic venous disorders. J Vasc Surg Venous Lymphat Disord 2021; 9(3): 568–584.

Hansrani

Riding

Seif

, et al. Pelvic vein incompetence and chronic pelvic pain: a case-control study. Bjog 2023; 130(11): 1355–1361.

10.

Champaneria

Shah

Moss

, et al. The relationship between pelvic vein incompetence and chronic pelvic pain in women: systematic reviews of diagnosis and treatment effectiveness. Health Technol Assess 2016; 20(5): 1–108.

11.

Gavrilov

Mishakina

Vasilyiev

, et al. Retrospective analysis of complications after gonadal vein interventions for pelvic venous disorder. J Surg Res 2023; 283: 249–258.

12.

Hansrani

Dhorat

McCollum

. Diagnosing of pelvic vein incompetence using minimally invasive ultrasound techniques. Vascular 2017; 25(3): 253–259.

13.

Riding

Pond

McCollum

, et al. Seeking consensus amongst UK-based interventional radiologists on the imaging diagnosis of pelvic vein incompetence in women with chronic pelvic pain: a modified Delphi study. Phlebology 2019; 34(7): 486–495.

14.

Corrion Otero

Ley Pozo

Zubicoa Ezpeleta

, et al. Contribution of Duplex color techniques in pelvic venous stasis. JEMU 1999; 20: 283–287.

15.

Bachar

Belenky

Greif

, et al. Initial experience with ovarian vein embolization for the treatment of chronic pelvic pain syndrome. Isr Med Assoc J 2003; 5(12): 843–846.

16.

Hansrani

Riding

Seif

, et al. Transvenous occlusion of incompetent pelvic veins to treat chronic pelvic pain in women: a randomised controlled trial. Bjog 2023; 130(11): 1362–1369.

17.

Zheng

Liu

. Ultrasound characteristics and risk factors of female patients with pelvic varicose veins and concomitant chronic pelvic pain. J Vasc Surg Venous Lymphat Disord 2025; 13(2): 102165.

18.

Halligan

Campbell

Bartram

, et al. Transvaginal ultrasound examination of women with and without pelvic venous congestion. Clin Radiol 2000; 55(12): 954–958.

19.

Creton

Hennequin

Kohler

, et al. Embolisation of symptomatic pelvic veins in women presenting with non-saphenous varicose veins of pelvic origin - three-year follow-up. Eur J Vasc Endovasc Surg 2007; 34(1): 112–117.

20.

Park

Lim

, et al. Diagnosis of pelvic congestion syndrome using transabdominal and transvaginal sonography. AJR Am J Roentgenol 2004; 182(3): 683–688.

21.

Barros

Perez

Zandonade

, et al. Evaluation of pelvic varicose veins using color Doppler ultrasound: comparison of results obtained with ultrasound of the lower limbs, transvaginal ultrasound, and phlebography. J Vasc Bras 2010; 9: 15–23.

22.

Whiteley

Dos Santos

Harrison

, et al. Transvaginal duplex ultrasonography appears to be the gold standard investigation for the haemodynamic evaluation of pelvic venous reflux in the ovarian and internal iliac veins in women. Phlebology 2015; 30(10): 706–713.

23.

Hainsworth

Eden

Ghosh

, et al. Detection and coil embolisation of pelvic vein incompetence for chronic pain in the female pelvis: a development phase economic analysis. Eur J Vasc Endovasc Surg. 2025 Mar; 69(3): 474–482.