Phase II trial of pulse dactinomycin as salvage therapy for failed low-risk gestational trophoblastic neoplasia

A Gynecologic Oncology Group study


  • The following member institutions participated in the study: University of Alabama at Birmingham, Duke University Medical Center, Abington Memorial Hospital, University of Mississippi Medical Center, University of Iowa Hospitals and Clinics, University of Texas Southwestern Medical Center at Dallas, Indiana University School of Medicine, University of California Medical Center at Irvine, State University of New York Downstate Medical Center, University of Kentucky, State University of New York at Stony Brook, Cooper Hospital/University Medical Center, Columbus Cancer Council, Women's Cancer Center, University of Oklahoma, and Case Western Reserve University.



The purpose of the study was to determine the activity and toxicity of pulse dactinomycin as salvage treatment of patients with low-risk gestational trophoblastic neoplasia (GTN) who failed methotrexate therapy.


Eligible patients had persistent/recurrent low-risk GTN defined by changes in serum human chorionic gonadotropin (hCG) levels (<10% fall over 3 consecutive weekly titers, >20% rise over the previous value, or a rise after attaining institutional normal [>5 mu/mL]); World Health Organization (WHO) score 2–6; Gynecologic Oncology Group (GOG) performance status 0–1; and previous treatment restricted to methotrexate. Dactinomycin administration was 1.25 mg/m2 intravenous (i.v.) every 2 weeks until documented complete response (CR) or treatment failure. CR was defined as an institutional normal serum hCG level sustained for ≥4 consecutive weeks; treatment failure was a <10% fall (3 assays over 4 weeks) or >20% rise (over previous value) in hCG serum level. Levels were monitored biweekly × 8 weeks beyond the first normal value, then monthly × 10.


Five of 44 enrolled patients were ineligible due to choriocarcinoma and normal pretreatment serum hCG level (2 each), no history of methotrexate (1), and 1 patient with documented phantom hCG syndrome was unevaluable. In all, 28 of 38 (74%) evaluable patients attained CR. The median number of cycles was 4 (range, 2–10). Severe toxicity was minimal, causing no patient to discontinue therapy. All treatment failures achieved a CR after receiving subsequent chemotherapy; 3 patients also underwent hysterectomy.


Pulse dactinomycin is an active regimen for patients withlow-risk GTN who fail previous methotrexate therapy. Cancer 2006. © 2006 American Cancer Society.

The activity of dactinomycin in gestational trophoblastic neoplasia (GTN) has been described in numerous studies, confirming its efficacy in both primary and recurrent/persistent GTN.1–8 At present, however, most low-risk patients are initially treated with single-agent methotrexate due to its known activity, low toxicity, and low cost.8–12

Between 10% to 31% of patients treated with single-agent methotrexate for low-risk disease (World Health Organization [WHO] score 0–6) will require a change in therapy due to tumor resistance.8–12 The majority of these recurrent patients are subsequently treated with either the 5-day regimen of dactinomycin (12 μg/kg/day, intravenous [i.v.] every 2 weeks), or with EMA/CO (etoposide, methotrexate, dactinomycin, cyclophosphamide, and vincristine). Although both regimens are associated with high cure rates, their attendant toxicity and cost represent a substantial burden for patients and the healthcare system.

Twiggs3 summarized the toxicity of both the 5-day and the pulse regimens of dactinomycin. Combining his data with Petrilli and Morrow's4 (both used 40 μg/kg), the rate of Grade 2 or 3 hematologic toxicity was 6 events in 82 cycles of pulse therapy. The corresponding rate as reported by Petrilli et al.,5 Goldstein et al.,1 and Ross et al.,2 using the 5-day regimen, was 14 events in at least 95 cycles of therapy (Ross et al. did not report the number of cycles). Gastrointestinal (GI) toxicity was greater in the pulse therapies (11/82 events vs. 0/95 events). An important issue to consider in these young, highly curable women is alopecia, which is universal and complete on the 5-day regimen, and which is significantly reduced on the pulse regimen. Petrilli et al.5 reported only 4 cases of Grade 1 (mild) alopecia using the pulse regimen. Twiggs3 also attempted a crude estimate of costs of therapy and found that the cost per cycle of pulse therapy was $74.75 compared with $342.25 for the 5-day regimen (1983 data). Finally, the pulse regimen requires only 1 visit every 2 weeks as opposed to most other regimens that involve at least weekly visits (if not 4–5 visits) every 2 weeks.

The magnitude of the comparable toxicities seen with EMA/CO is apparent from the report by Quinn et al.13 They reported a rate of Grade 3 and 4 toxicity events in 65 high-risk patients treated with EMA/CO as neutropenia, 41%; thrombocytopenia, 17%; alopecia, 40%; and nausea and vomiting, 22%.

In 1987, Petrilli et al.5 reported on a Phase II Gynecologic Oncology Group (GOG) study using pulse dactinomycin (1.25 mg/m2, i.v. every 2 weeks) in previously untreated patients with nonmetastatic GTN.5 Ninety-four percent of those patients achieved remission after receiving a median of 4 cycles. The 2 treatment failures were subsequently cured with alternative therapies. Five percent of all cycles were associated with Grade 3 GI toxicity; however, this and all reports of dactinomycin predate the introduction of 5-HT3 antagonists into antiemetic regimens. There were no Grade 3 or 4 hematologic toxicities reported.

Among 3 other reports of pulse dactinomycin as first-line therapy in nonmetastatic GTN, the combined primary remission rate was 92%, which is similar to that reported for the 5-day regimen.3, 5, 7

Based on these data, the current study was undertaken to determine the efficacy and toxicity of pulse dactinomycin as salvage therapy in patients who failed 1 previous methotrexate-containing regimen for low-risk GTN.


Eligible patients were those who had a histologically confirmed previous hydatidiform molar pregnancy (choriocarcinoma and placental site trophoblastic tumor were ineligible) and who had failed first-line therapy with methotrexate (with or without folinic acid). Also required were a WHO score of 2–6 (Table 1) and a GOG performance status of 0–1. The serum human chorionic gonadotropin (hCG) entry criteria included any of the following: >20% rise over a previous value, <10% fall over 3 consecutive weekly levels, or any rise in serum hCG after attaining institutional normal. Pretreatment eligibility also included white blood cell (WBC) ≥3000/mm3, platelets ≥100,000/mm3, creatinine ≤1.5 mg%, bilirubin ≤1.5 times institutional upper limit of normal, and SGOT (serum glutamic oxaloacetic transaminase) and alkaline phosphatase ≤3 times the upper limit of institutional normal. Patients were required to use an accepted method of contraception (oral contraceptives or male/female sterilization). Patients with metastases to the liver, spleen, brain, kidney, or GI tract, or with more than 8 identified metastatic lesions, were ineligible. Additionally, patients whose antecedent pregnancy terminated more than 1 year before entry, or those with invasive malignancies other than nonmelanoma skin cancer, were ineligible. Patients with evidence of other cancer present within the last 5 years, or whose previous cancer treatment contraindicated protocol therapy, were also ineligible. Participating institutions obtained approval from their Institutional Review Boards before entering any patient in the study and all patients provided written informed consent consistent with all federal, state, and local requirements before receiving protocol therapy.

Table 1. WHO Scoring Schema
Patient characteristicScore*
  • hCG, human chorionic gonadotropin; GI, gastrointestinal; WHO, World Health Organization.

  • *

    The total WHO score for a patient is obtained by adding the individual scores corresponding to each observed characteristic.

Age at diagnosis (y)<40≥40  
Antecedent pregnancyMoleAbortionTerm 
Interval (mos between end of antecedent pregnancy and study registration)<44–67–12≥12
Serum hCG (IU/L) at study entry<103103–104104–105≥105
Largest tumor (including uterine tumor)<33–5 cm≥5 cm 
Site of metastasesNone, lung, or vaginaSpleen/kidneyGI tract/liverBrain
Number of metastases01–45–8>8
Prior chemotherapyNoneSingle≥2 Drugs

Dactinomycin was administered i.v. at a dose of 1.25 mg/m2, to a maximum single dose of 2.0 mg, every 2 weeks. The use of an antiemetic regimen was left up to the individual treating physician; however, a 5-HT3 antagonist with dexamethasone was recommended. Therapy was to continue until evidence of complete response (CR) or treatment failure, or (for patients whose serum hCG level reached institutional normal) for an additional 2 cycles. CR was defined as an institutional normal serum hCG sustained for ≥4 consecutive weeks; cure was defined as maintenance of the institutional normal serum hCG for 1 year after completion of protocol therapy. Treatment failure was a >20% rise over a previous value or <10% fall over 3 consecutive biweekly levels. Recurrence was any rise in serum hCG after attaining institutional normal. Progression was defined as the development of new metastases. Serum hCG levels were obtained every 2 weeks during treatment and for 2 months after attaining institutional normal, then monthly for 10 additional months. Patients were evaluated for response and toxicity every 2 weeks. Toxicity was graded according to the Common Toxicity Criteria, v. 2.0.

An optimal 2-stage sampling design was utilized for this Phase II study.14 A CR rate of 80% or greater was deemed clinically significant, whereas a CR rate of 60% or less was declared insufficient to warrant further investigation. The target accrual for the first stage was 15 eligible patients among whom at least 10 complete responses were required to continue to the second stage of accrual. The total target accrual was 35 patients. Type I and type II error were set at 0.10. The estimated chance of early termination after completing the first stage was 60%.14 Whereas the initial design of this study targeted a fixed sample size at each stage, it is difficult to attain a specific sample size within the setting of a multisite trial. Subsequently, a flexible design utilizing the same design parameters as outlined above was applied to provide decision rules at each stage. Allowing for a range of attained sample sizes and the same design parameters, the decision rule at the second stage for a sample size of 38 is to reject the null hypothesis of a response rate equal to 60% when 27 or more responses are observed.15 One-sided confidence limits conditioned on early stopping rules were constructed.16 Logistic regression was used to evaluate the relation between hCG level and WHO score at study entry and the odds of response. Given the small sample size, the conclusions in these types of analyses should be interpreted with caution due to the limited power of detecting associations within the confines of this study.


Between 1999 and 2004, 44 patients were entered into this study (Table 2). Upon central review, 4 patients were excluded due to a choriocarcinoma in the original tumor (n = 2) and normal serum hCG at the time of registration (n = 2). In addition, 1 patient had no history of methotrexate therapy and 1 patient was inevaluable for response because she had phantom hCG syndrome documented after initiation of protocol treatment.17 Patient characteristics are described in Table 2. The WHO score at commencement of first-line therapy is listed in Table 3. All patients had WHO scores of 6 or less when first-line therapy was initiated. All patients received prior methotrexate therapy, and although the regimen was not restricted (weekly vs. 5-day), all but 2 were previously treated on a weekly regimen. All patients had negative chest X-rays for lung metastases, although 1 was positive for lung metastases on chest CT (which was allowed pretreatment).

Table 2. Patient Characteristics and Treatment Information
CharacteristicNo.PercentNo. complete responses/No. evaluable for response
  • GTD, gestational trophoblastic disease; CXR, chest x-ray; hCG, human chorionic gonadotropin.

  • *

    From antecedent pregnancy to study registration.

Age at registration, y
Antecedent pregnancy
Mole class
 Complete mole3282 
 Partial mole718 
Interval* (mo)
WHO score at study entry
Previous GTD717 
Abnormal pelvic ultrasound1846 
Largest tumor including uterine tumor
 <3 cm328221/31
 3–5 cm6156/6
 >5 cm131/1
No metastases identified by CXR39100 
Prior single-agent chemotherapy39100 
hCG at registration (IU/L)
Treatment information
 Median number of cycles (range)4 (2–10) 
 Median total dose (range)9 mg (2–21) 
Table 3. WHO Score at Start of First-Line Therapy
WHO ScoreNo.Percent

Of the 38 patients evaluable for response, 28 (74%, 90% confidence interval [CI]: 60%–85%) attained a normal institutional serum hCG level that was sustained for at least 1 month. Documentation of sustained normal hCG levels for 12 months was available in 21 of these 28 patients. There was insufficient documentation of a long-term response in the other 7 patients who were lost to follow-up: 11 months (2); 8 months (1); 6 months (1); 5 months (1); 3.5 months (1); 2 months (1). Recognizing the small sample size, there was no statistically significant correlation between WHO score and patient response to this regimen (Table 2). The median serum hCG at study entry was 290 (6–34,190). The association between hCG and response was evaluated by a logistic regression model. The estimate of the CR odds ratio for log hCG was 1.025 (95% CI: 0.75–1.41), P = .88. Responses for categorized hCG can be found in Table 2. Additionally, using a cut-off of 100, there were 9 complete responses among 11 patients with serum hCG levels <100 IU/L and 19 complete responses among 27 patients with hCG levels >100 IU/L. There is insufficient information to evaluate response by previous methotrexate regimen.

Therapy was not discontinued in any patient due to toxicity. The treatment was well tolerated, with only 4 patients (10%) experiencing Grade 3 or 4 neutropenia and 2 patients (5%) experiencing Grade 3 nausea and vomiting out of 39 patients evaluated for toxicity (Table 4). One patient experienced Grade 3 toxicity from drug extravasation. Eleven patients experienced Grade 1 alopecia and 5 experienced Grade 2 alopecia. All other dermatologic toxicity was limited to Grade 1 and included acneiform rash (4), flushing (1), pigmentation changes (1), pruritus (1), burning sensation during i.v. (1), dry skin (1), and bruising (1). These effects were reported in 6 patients.

Table 4. Patients with Grade 3 or 4* Toxicity (n = 39)
ToxicityGrade 3Grade 4
  • *

    Number of patients experiencing maximum observed grade of toxicity.


Each of the 10 patients who did not respond to protocol treatment achieved a CR with additional salvage chemotherapy (Table 5). Two of the 10 patients were continued on this regimen by their treating physicians despite meeting the protocol criteria for treatment failure (>20% rise over previous value). They both eventually achieved a CR with no additional change of therapy and are detailed. In 1 patient, her hCG went from 16 to 22 in 2 weeks (38% increase). The patient was removed from the study, yet continued on the same regimen as per her treating physician. She attained an institutional normal hCG after a further 4 cycles of dactinomycin. The other patient achieved an hCG level of 5 after 10 cycles of dactinomycin, then subsequently rose to 7 (at 2 weeks) and 39 (at 4 weeks). The patient was removed from the study, yet continued on the same regimen as per her treating physician. She attained an institutional normal hCG within 4 weeks. Note that 3 patients were treated by both chemotherapy and hysterectomy.

Table 5. Salvage Regimens (n = 10)
RegimenChemotherapy aloneChemotherapy and hysterectomy
  • *

    ACT-D followed by methotrexate in 1 patient.

  • EMACO × 6 followed by EMA.

Dactinomycin (ACT-D)
Ifosfamide/cisplatin/etoposide (ICE)1 
Etoposide/methotrexate/actinomcyin-D (EMA)11
Etoposide/methotrexate/actinomcyin-D/cyclophosphamide/vincristine (EMACO)11

Six women had 8 subsequent pregnancies; of these, there were 6 full-term deliveries and 2 spontaneous abortions.


McNeish et al.8 reported on a single institution's experience using 5-day dactinomycin to treat patients with GTN who failed previous therapy with methotrexate. Of 161 patients with a Charing Cross score ≤8 (equivalent to a WHO score ≤6), 67 were treated with dactinomycin and the remission rate was 87% (95% CI: 76%–94%). It is difficult to determine if the patient populations in these 2 studies are similar given the differences in eligibility criteria (twice weekly hCG levels) and the fact that 2% of patients stopped methotrexate due to toxicity as opposed to drug resistance. In that study 53 (80%) patients had hCG levels <100 IU/L, whereas in our study only 12 (31%) patients had levels that low. Further, in these patients our response rate was 83%, which is not dissimilar to their reported overall results. Interestingly, similar to our analysis, McNeish et al.8 found no correlation between dactinomycin failure and hCG level (> or <100 IU/L), and in fact suggested that dactinomycin was reasonable to use in patients with an hCG up to 500 IU/L.

The potential for cure is not the sole consideration when selecting second-line therapy for failed low-risk GTN. Although cure can eventually be attained in nearly 100% of these patients,8 other issues of importance include toxicity, cost, convenience, patient compliance, and preference. Whereas other regimens including EMA/CO may be potentially more effective, they confer substantially greater (primarily hematologic) toxicity, cost, and inconvenience when compared with pulse dactinomycin.8 Alopecia and acneiform rash are troubling toxicities for these young women. Hair loss appears to be reported much less often (of 40 patients, 5 reported pronounced hair loss and 12 reported mild hair thinning) with this regimen compared with 100% reported using the 5-day regimen or etoposide-containing regimens. Similarly, in our series only 3 patients experienced Grade 1 acne form rash and, whereas this toxicity is not well characterized in the literature due to a lack of prospective analysis, it is our impression that it occurred less frequently and less severely than in the 5-day regimen. Also, the use of etoposide (arguably the most effective single drug to treat GTN) confers an increased risk of acute myelocytic leukemia, colon cancer, and premature menopause,18 important issues to consider in a young, highly curable patient population.

There have been only a few prospective multisite studies in GTN, and none in failed low-risk patients, making this study unique. However, this study demonstrates the need for more precise definitions of treatment failure. The definitions of plateau and rise used in this study were previously published by the GOG and others3, 5, 7, 11, 19, 20 and, whereas these definitions are used by many investigators, they are imperfect, particularly for patients with low hCG levels. For example, in 1 week 1 patient experienced a rise in hCG from 10 IU/L to 13, which in this protocol was considered a treatment failure. But this situation carries clearly different implications than one in which there is a serum hCG rise from 500 IU/L to 650 IU/L, even though both represent a 30% increase. The significance of this problem with low hCG levels has been previously identified21 and is highlighted in Table 5, wherein 2 patients who were defined as treatment failures in this study were continued on the same regimen by their physician and both ultimately achieved a CR. There are inherent problems associated with the use of a single hCG value, not the least of which is laboratory error, which can lead to an incorrect clinical decision. This broad issue may account for the significant differences in efficacy observed between prospective multi- and single-site studies where a well-defined protocol, rather than clinical judgment and experience, may influence decisions concerning therapy.

Dobson et al.22 reported on etoposide and dactinomycin (EA) to treat 35 low-risk GTN patients who had failed prior methotrexate therapy. Whereas the remission rate of 97% (95% CI: 85%–99%) was very high, toxicity in that study was substantial, including 100% of patients with total alopecia, 57% with Grade 3–4 neutropenia, 28% with Grade 2–4 nausea and vomiting, and 37% with stomatitis.

Kohorn23 suggested that lack of response to pulse dactinomycin may be more related to scheduling than to drug resistance. He reported that 4 of 5 patients who failed pulse dactinomycin therapy were salvaged with 5-day dactinomycin, suggesting the importance of drug exposure. However, no strict definition of a plateau or rise in serum hCG levels was given, and in only 1 of the patients described is it apparent that she met the definition of treatment failure in this protocol.

Only in a randomized trial could the best second-line regimen for relapsed low-risk GTN patients be determined. It would be difficult to mount such a study given that it would require at least 260 patients to test the superiority of 5-day dactinomycin to pulse dactinomycin (α = 0.05, β = 0.2) using the results obtained in this study and those quoted by McNeish et al.8 The belief of some investigators, that the 5-day regimen is more efficacious, may be based on single-site institutional experiences consisting of mostly small numbers of patients. Although differences in efficacy are equivocal, it would appear that toxicity, cost, and convenience are dissimilar between these regimens. With a salvage rate of virtually 100%, it does not seem unreasonable to use drugs/regimens that are of proven effectiveness, with minimal toxicity and cost, and whose administration is convenient for the patient and healthcare team. If 25% of low-risk GTN patients fail methotrexate and 25% of those fail second-line pulse dactinomycin, then only 6% of patients treated with first-line methotrexate will eventually require either surgery or combination chemotherapy. Therefore, it is difficult to advocate the use of combination chemotherapy (particularly regimens containing etoposide) as second-line therapy for patients who fail methotrexate or have persistent low-risk GTN. One significant advantage of pulse dactinomycin over 5-day dactinomycin is the lower risk of extravasation and potential slough after intravenous administration.

There are obvious inherent weaknesses in this study. Its size, whereas reasonable for a prospective multisite study, is relatively small compared with the series of single-site studies published over many years.8–10, 12 Whereas every patient was followed for at least 8 weeks after attaining a normal hCG, loss to long-term follow-up was problematic. However, it can be argued that multisite studies are likely more externally generalizable, and may therefore be more applicable to the typical patient who is diagnosed with this cancer worldwide.

Until the emergence of definitive evidence of a superior regimen (high efficacy with minimal toxicity, low cost, and convenience), pulse dactinomycin represents a rational treatment option for patients with low-risk GTN who fail initial treatment with methotrexate. These data suggest that this regimen is worthy of further investigation.