Uterine artery pulsatility index improves prediction of methotrexate resistance in women with gestational trophoblastic neoplasia with FIGO score 5–6


Correspondence: Dr R Agarwal and Professor MJ Seckl, Department of Medical Oncology, Charing Cross Hospital Campus, Imperial College London, Fulham Palace Road, London W6 8RF, UK. Email roshan.agarwal@imperial.ac.uk; m.seckl@imperial.ac.uk



The Uterine Artery Pulsatility Index (UAPI) is an ultrasound measure of tumour vascularity. In this study, we hypothesised that a UAPI ≤1 (high vascularity) would identify women with gestational trophoblastic neoplasia (GTN) at increased risk of resistance to first-line single-agent methotrexate (MTX-R).


Single-centre cohort study.


Charing Cross Hospital, a UK national centre for the treatment of trophoblastic disease.


All women with a GTN FIGO score 5–6 treated with methotrexate (n = 92), between 1999 and 2011, at Charing Cross Hospital.


UAPI was measured before the start of chemotherapy, and women were monitored for the development of MTX-R.

Main outcome measures

Frequency of MTX-R in women with UAPI ≤1 compared with UAPI >1.


UAPI was measured before chemotherapy in 73 of 92 women with GTN FIGO score 5–6. UAPI ≤1 predicted MTX-R independent of the FIGO score (hazard ratio 2.9, P = 0.04), with an absolute risk of MTX-R in women with a UAPI ≤1 of 67% (95% CI 53–79%) compared with 42% (95% CI 24–61%) with a UAPI >1 (P = 0.036).


Our results suggest UAPI is an independent predictor of MTX-R in women with FIGO 5–6 GTN.


Gestational trophoblastic neoplasia (GTN) is one of the few malignancies that can be cured with the use of cytotoxic chemotherapy alone. The majority of women develop GTN following a molar pregnancy, and are diagnosed based on rising serum human chorionic gonadotrophin (hCG) levels following termination of the antecedent pregnancy. The choice of chemotherapy used to treat these women is based on their International Federation of Obstetrics and Gynecology (FIGO) prognostic score.[1] Women with a FIGO score 0–6 (low-risk) receive single-agent methotrexate as first-line therapy.[1] However, approximately 45% of women with low-risk GTN develop resistance to methotrexate (MTX-R), and most of these require a switch to multi-agent etoposide, methotrexate, actinomycin D/cyclophosphamide, vincristine (EMA/CO) chemotherapy.[1]

MTX-R is a major source psychological distress for women with GTN, and also results in an increase in the overall treatment duration, and delays the ability of women to conceive. Better prediction of the risk of MTX-R is therefore needed.

The uterine artery pulsatility index (UAPI) is a Doppler ultrasound measure of tumour vascularity, and we have previously shown that a UAPI ≤1 (high vascularity) is an independent predictor of MTX-R in women with low-risk disease. The predicted risk of MTX-R in women with a UAPI ≤1 and a FIGO score of 5–6 was >75%, suggesting that these women might benefit from initial EMA/CO therapy.[2, 3] However, a significant limitation of our previous two studies was that women with FIGO 5–6 GTN constituted only a small proportion (10%) of the overall patient cohorts (16/164 and 24/239 women).[2, 3] The current study, therefore, focused on women with FIGO 5–6 GTN, treated over a 12-year period, to determine more precisely the predictive value of UAPI in this group of women.


The Trophoblastic Disease Centre's electronic patient database at Charing Cross Hospital—one of two UK national centres for the treatment of GTN—was searched for all women with GTN FIGO 5–6, treated with methotrexate over a 12-year period (January 1999 to June 2011; Figure 1). The women included in this series had all undergone uterine evacuation, met the Charing Cross Hospital guidelines for treatment and had follow up of at least 1 year after the completion of therapy to monitor for recurrence.

Figure 1.

Flow diagram of women.

Chemotherapy after a molar pregnancy was instituted if: the woman had heavy vaginal bleeding, a rising hCG level over a 2- to 4-week period, a plateau of 4 weeks or longer in hCG levels, or a serum hCG level 4 weeks after evacuation >20 000 IU/l. Women meeting these criteria were assessed clinically, and staged with a pelvic Doppler ultrasound scan, chest X-ray and repeat serum hCG. At the time of baseline Doppler ultrasound scan the UAPI was also assessed as previously described.[2, 3] The lowest of the left and right UAPIs in a woman was used for the correlation with MTX-R. The co-efficient of variation of measurement of UAPI was <10%.[2, 3]

The standard first-line therapy for women with a FIGO score of 6 or lower was the Charing Cross Hospital Methotrexate and Folinic acid regimen (MTX/FA).[1] A small number of women with FIGO 5–6 and additional adverse characteristics such as heavy bleeding, hCG levels over 250,000 IU/l or a contraindication to intramuscular therapy were commenced on first-line treatment with either D1-5 dactinomycin or the EMA/CO combination regimen. These women were excluded from this study.[4]

Resistance to first-line methotrexate (MTX-R) was defined as a plateau or rise in two consecutive hCG measurements. Second-line chemotherapy following MTX-R was determined by the serum hCG at the time of resistance. Women with hCG levels ≤300 IU/l received single-agent dactinomycin and those with hCG levels >300 IU/l were commenced on EMA/CO. Women developing second-line resistance to dactinomycin were also switched to EMA/CO. Women who changed chemotherapy because of drug side effects (serositis or allergy) were not considered to be resistant.

Statistical analysis

A UAPI cut-off of ≤1 and >1 was used based on the results from previous studies. Logistic regression with forward selection was used to determine the utility of UAPI independent of the FIGO score for predicting MTX-R. A two-sided α of 5% was used to denote statistical significance. The study had 86% power with 73 women to detect a difference in MTX-R rates of 35% versus 70% in women with FIGO 5–6 and UAPI ≤1 and >1, respectively. MTX-R rates for the power calculation were based on results from our previous studies. Statistical analysis was performed using SPSS V17 (SPSS, Chicago, IL, USA).


Between January 1999 and June 2011, 92 women with GTN FIGO score 5–6 were treated with MTX/FA (Figure 1). Of these, 73 women were assessable for both UAPI and MTX-R, and were comparable to the overall cohort with respect to baseline characteristics (P > 0.05, data not shown). The median UAPI was 0.74 (range 0.1–2.9). The median age of women was 33 years (range 16–53), and median hCG was 63 707 IU/l (range 175–1 217 592 IU/l). Two women had pulmonary metastases at the time of diagnosis.

Overall, 43 women developed MTX-R (58.9%; 95% CI 47–69%); they were switched to second-line EMA/CO (n = 33) or dactinomycin (n = 10) chemotherapy, as described (see 'Design'). Four of the ten women switched to dactinomycin, required third-line EMA/CO for resistance. All women (n = 73) were cured.

The risk of MTX-R in women with UAPI ≤1 and UAPI > 1 was 67% (33 of 49; 95% CI 53–79%) versus 42% (10 of 24; 95% CI 24–61%, = 0.036), respectively (Figure 1, Table 1). A UAPI ≤1 predicted resistance to methotrexate, independent of the FIGO score (odds ratio 2.9, P = 0.04), in logistic regression analyses. Rates of MTX-R following stratification by FIGO score were 43% (95% CI 21–68%) for women with a FIGO score of 5 and UAPI >1, compared with 75% (95% CI 50–90%) in women with FIGO score 6 and UAPI ≤1 (Table 1).

Table 1. MTX-R by FIGO score and UAPI
FIGO scoreUAPI >1 (n = 24)UAPI ≤1 (n = 49)
n %95% CI n n %95% CI n


Main findings

The results from this study validate our previous work and show that increased uterine blood flow (UAPI ≤1) is an independent predictor of MTX-R risk, in particular in women with GTN FIGO 5–6, and redresses the issue of the limited number of women with FIGO score 5–6 in our previous studies. There is clearly scope to shorten the duration of treatment and to reduce psychological distress for the 45% of women who develop MTX-R and require more potent and toxic EMA/CO chemotherapy, by identifying a subset of these women before chemotherapy.

Strengths and weaknesses

The odds ratio for MTX-R in the current study was similar to our previously published results, and indicates that the risk of MTX-R associated with increased angiogenesis is constant, and independent of risk factors associated with disease burden and metastatic potential of women with low-risk GTN (FIGO 0–6). Our current results suggest that upfront EMA/CO chemotherapy should also be considered for women with GTN with FIGO scores 5–6, and a UAPI ≤1. This could be achieved by incorporation of UAPI into the FIGO prognostic scoring system by adding 1–2 points for a UAPI ≤1.

The biological processes under-pinning the mechanism of MTX-R in GTN in general and angiogenesis are unclear. In other tumours receptor tyrosine kinase signalling pathways associated with fibroblast, platelet-derived and vascular endothelial growth factors have been implicated in both angiogenesis and drug resistance, and we are currently exploring the association between these growth factors, UAPI and MTX-R in GTN.

We have previously shown that a baseline hCG level >400 000 IU/l also appears to be an adverse predictive factor for MTX-R and is used for treatment selection for high-risk chemotherapy at Charing Cross Hospital.[4]

In the absence of markers for MTX-R with 100% specificity, given the 100% salvage rates of women with MTX-R, there is a strong rationale for not over-treating women and exposing them to the risks of an earlier menopause and of the secondary tumours that are related to EMA/CO but absent with MTX/FA.[5, 6] The clinical value of UAPI may therefore be greatest in treatment selection when there is strong preference among women for minimising the need for changes in treatment, acceptance of the greater toxicity of EMA/CO, and a desire to become pregnant again as soon as possible.

One limitation of our study is that the patient cohort is derived from a single institution (Charing Cross Hospital), and as such there is the potential for selection bias common to unblinded non-randomised studies. The women, however, represent a consecutive series of low-risk gestational trophoblastic neoplasia, and Charing Cross Hospital is one of the two national referral centres. As a consequence, our results are unlikely to be subject to referral/selection bias compared with centres without national registration and treatment policies.


Our results address the issue of limited numbers of women with FIGO 5–6 disease in our previous studies, and suggest that consideration be given to the incorporation of UAPI into the FIGO prognostic scoring system, with the addition of 1–2 points for a UAPI ≤1, subject to further validation.


A UAPI ≤1 predicts an increased risk of MTX-R in women with low-risk GTN, and may help in patient stratification for first-line chemotherapy. A prospective study to confirm these findings has been agreed with the Gynaecological Oncology Group in the context of a planned randomised study of chemotherapy regimens in low-risk GTN (FIGO 0–6).

Disclosure of interests

The authors declare that there are no conflicts of interest.

Contribution to authorship

AS-L and HM are joint first authors with responsibility for data collection, analysis and interpretation. RF performed data collection; DS performed administration and data collection and NS and RH analysed and interpreted the data. PS was the medical oncology consultant responsible for patient care and date interpretation and AL was the radiology consultant responsible for UAPI data. MS was – medical oncology consultant responsible for patient care and data interpretation and RA – was the medical oncology consultant who conceived the idea of study, and performed data analysis and interpretation. All authors were involved in the writing of the manuscript.

Details of ethics approval

This study was carried out with anonymised patient records and hence informed consent was not needed.


The GTD service at Charing Cross Hospital is funded by the National Commissioning Group and Department of Health. We would also like to acknowledge support from the Imperial College Experimental Cancer Medicine Centre (ECMC) and Biomedical Research Centre (BRC) grants from Cancer Research UK, NIHR and the UK Department of Health.


RA is funded by a Cancer Research UK Clinician Scientist Fellowship. MJS is supported by the Department of Health National Commissioning Group, and would like to acknowledge funding from the Cancer Research UK and Department of Health Imperial Experimental Cancer Medicine Centre and Biomedical Research Centre grants.