A phase 2 study of platinum and gemcitabine in patients with advanced salivary gland cancer

A Trial of the NCIC Clinical Trials Group

Abstract

BACKGROUND:

Salivary gland cancers are rare, histologically diverse, and varied in their biologic behavior and responsiveness to systemic therapy. To the authors' knowledge, there currently is no standard chemotherapy for these tumors, but cisplatin-based regimens are often used. This phase 2 trial evaluated the combination of gemcitabine with cisplatin (carboplatin in those with protocol-defined contraindications to cisplatin).

METHODS:

Fit, consenting adult patients had advanced, metastatic, or locoregionally recurrent salivary gland cancer (any histologic subtype) that was not suitable for radiation or surgery. Therapy was comprised of gemcitabine at a dose of 1000 mg/m² administered intravenously on Days 1 and 8, and cisplatin at a dose of 70 mg/m² on Day 2, of a 21-day cycle. If carboplatin was substituted, it was administered on Day 1, targeted to an area under the concentration-time curve of 5 mg/mL/s. Response was assessed every 2 cycles according to Response Evaluation Criteria In Solid Tumors. Patients received up to 6 cycles. The primary endpoint was objective response. A 2-stage design was used, with a response rate of 45% required to declare the regimen active.

RESULTS:

Thirty-three eligible patients were enrolled, of whom 30 were response evaluable. Eight objective responses were observed (1 complete and 7 partial) for a response rate of 24% (95% confidence interval, 11-42%), with responses observed in all histologic subtypes. Toxicity was within that expected for this combination.

CONCLUSIONS:

This regimen did not meet the predefined criteria to be declared active in advanced salivary gland cancers. Enrollment of patients with these rare cancers into well-designed clinical trials remains an urgent priority. Cancer 2010. © 2010 American Cancer Society.

Malignant tumors of salivary gland origin are rare cancers, accounting for <5% of all malignancies of the head and neck. They are among the most diverse malignancies, with 24 different histologic subtypes,1 a wide spectrum of biologic behavior, and variable responsiveness to systemic therapies. Therapy of localized disease generally is comprised of surgery, radiotherapy, or both. Once the disease is metastatic, therapy is generally palliative, although some patients may have a prolonged, indolent course. For example, in case series of adenoid cystic carcinoma (ACC) of salivary gland origin, the median duration of survival with metastatic disease was approximately 3 years,2, 3 with up to 10% of patients surviving for >10 years.4, 5 Chemotherapy is reserved for the palliative therapy for metastatic disease or for locoregional disease for which further surgery or radiation is not possible. However, because many patients may have indolent disease and few disease-related symptoms, chemotherapy should be reserved for those patients who are symptomatic, or at risk of developing symptoms due to disease progression; consideration should also be given first to local therapies (eg, metastatectomy of isolated lung metastases in those with indolent disease) before resorting to chemotherapy.

Given their rarity, there are limited data regarding the role of chemotherapy in the palliative treatment of advanced salivary gland cancers.6, 7 To the best of our knowledge, there is no accepted standard chemotherapy regimen, but cisplatin-based combination chemotherapy, usually in conjunction with an anthracycline, is commonly used.6 The combination of cisplatin and gemcitabine is now an accepted regimen for the treatment of patients with advanced nonsmall cell lung cancer8 and urothelial cancer.9 This study was designed to evaluate the antitumor efficacy of this combination in the treatment of patients with advanced salivary gland cancers.

MATERIALS AND METHODS

Patient Eligibility

This trial enrolled adult patients with a malignant salivary gland tumor that was either metastatic, locoregionally recurrent after prior curative surgery and/or radiotherapy, or locally advanced and not suitable for surgery or radiotherapy. Any histology was permitted, but those patients with indolent histologies (such as acinic cell carcinoma, or low–grade mucoepidermoid carcinoma [MEC]) were required to be symptomatic or at imminent risk of developing disease-related symptoms. One prior non-platinum–containing, non-gemcitabine–containing regimen was permitted for recurrent or metastatic disease; 1 platinum-based adjuvant therapy regimen was permitted if completed >12 months before enrollment and if there was no residual neurotoxicity or ototoxicity of greater than grade 1. At least 28 days must have elapsed from prior radiotherapy, and 21 days from prior major surgery.

Patients were required to have an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2, a life expectancy of at least 12 weeks, unidimensionally measurable disease, and adequate end-organ function (granulocyte count ≥1.5 × 10⁹/L, platelet count ≥100 × 10⁹/L, aspartate aminotransferase and alanine aminotransferase ≤3 × the upper limit of normal, and calculated creatinine clearance of ≥60 mL/minute for cisplatin and between 30 and 59 mL/minute for carboplatin). Patients were excluded if they had brain metastases, serious comorbid conditions, were pregnant or lactating (or refused adequate contraception if of childbearing potential), or had a history of another malignancy (except for adequately treated non-melanomatous skin cancer or in situ cancer of the cervix) within the 5 years before enrollment.

The study was approved by the research ethics boards of the participating institutions, and all patients provided written informed consent. The study was conducted in accordance of Good Clinical Practice guidelines. The NCIC Clinical Trials Group (CTG) maintained the database and performed all analyses.

Treatment

All patients received gemcitabine at a dose of 1000 mg/m² intravenously (iv) on Days 1 and 8 of a 21-day cycle. Cisplatin, at a dose of 70 mg/m² iv with standard pre- and posthydration, was administered on Day 2. Carboplatin, administered iv on Day 1 at an area under the concentration-time curve (AUC) of 5 mg/mL/s, was substituted for cisplatin in the case of neurotoxicity or ototoxicity of ≥grade 2, or in the case of renal dysfunction (calculated creatinine clearance of between 30 and 59 mL/min). Standard antiemetics were administered.

Day 1 of therapy was delayed if the absolute neutrophil count was ≤1.5 × 10⁹/L or the platelet count was ≤100 × 10⁹/L. Day 8 gemcitabine was administered at full dose if the absolute neutrophil count was ≥1.0 × 10⁹/L and the platelet count was ≥100 × 10⁹/L; it was reduced by 25% if the absolute neutrophil count was 0.5 to 0.99 × 10⁹/L and/or the platelet count was 50 to 99 × 10⁹/L, and was omitted if either count was below these latter values. If the nadir absolute neutrophil count was <1.0 × 10⁹/L for ≥7 days, if the patient had febrile neutropenia, or if the nadir platelet count was <10 × 10⁹/L or the patient had thrombocytopenia-induced bleeding, all subsequent gemcitabine doses were reduced by 25% and, for those patients receiving carboplatin, the dose was reduced to AUC = 4. Treatment was held until resolution to ≤grade 2 toxicity for grade 3 nonhematologic toxicity. Carboplatin (AUC = 5) was substituted for cisplatin if grade 2 or greater sensorineural hearing loss or peripheral neuropathy was observed. Protocol therapy was permanently discontinued in the case of grade 4 nonhematologic toxicity, if the patient developed renal insufficiency with a calculated creatinine clearance of <30 mL/minute, if interstitial pneumonitis considered possibly related to gemcitabine developed, or if a cycle was delayed for >21 days due to unresolved toxicities.

Patients could receive up to 6 cycles of therapy. Protocol therapy was discontinued in the setting of disease progression, serious or unmanageable toxicity, or at the patient's request.

Pretreatment Evaluation and Evaluation on Therapy

Patients were assessed within 7 days of registration with a history and physical examination (including weight and ECOG PS), complete blood count (CBC) and serum biochemistry, urinalysis, and pregnancy test (if applicable). Computed tomography scans were performed as necessary to document sites of disease within 28 days before registration. While receiving therapy, a physical examination was performed on Day 1 of every cycle. A CBC was performed weekly during Cycles 1 and 2, and on Days 1 and 8 of subsequent cycles, whereas biochemistry was performed on Day 1 of every cycle. Radiographic studies to document tumor response were performed every second cycle. Response was determined according to Response Evaluation Criteria In Solid Tumors (RECIST),10 and toxicity was assessed using the National Cancer Institute Common Toxicity Criteria (version 2.0).

Objectives and Statistical Considerations

The primary objective of the current study was to estimate the activity of this combination chemotherapy among patients with malignant salivary gland tumors, as measured by the response rate. Secondary objectives included the determination of the duration of response and overall survival, and an assessment of the toxicity of this combination regimen in this patient population. Overall survival was defined as the time from the first day of treatment to death from any cause. Response duration was measured from the date of the first objective assessment of response to the date of first documentation of disease progression.

A 2-stage design was used for calculation of sample size.11 In stage 1, 11 response-evaluable patients were to be entered, and if at least 3 responses were observed, then an additional 23 response-evaluable patients were to be accrued, for a total of 34 patients. The combination would be declared active if ≥13 responses were observed. This tests the null hypothesis that the response rate is 25%, versus the alternate hypothesis that the response rate is 45%, with a significance level of α = 0.05 and β = 0.20.

RESULTS

Patients

Between October 2003 and October 2007, 34 patients were enrolled to this trial at 4 centers. One patient was registered, but before therapy was determined to have brain metastases, and was deemed ineligible. The characteristics of the 33 eligible patients are summarized in Table 1. The majority of the patients had either ACC (n = 10) or adenocarcinoma (n = 8) histology.

Table 1. Patient Characteristics

Characteristic	No. of Patients
ECOG indicates Eastern Cooperative Oncology Group. a Salivary duct carcinoma (1 patient), adenosquamous carcinoma (1 patient), and undifferentiated (4 patients).
Eligible patients	33
Female	14
Median age (range), y	58 (33-81)
ECOG performance status
0	14
1	17
2	2
Prior chemotherapy
Adjuvant	2
For metastatic disease	3
0 prior regimens	28
1 prior regimens	4
2 prior regimens	1
Prior radiotherapy	28
Histology
Adenoid cystic carcinoma	10
Adenocarcinoma	8
Acinic cell carcinoma	5
Mucoepidermoid carcinoma	4
Other histologiesa	6
≥3 Sites of disease	15

Treatment Received

Of the 33 eligible patients, 1 had interval deterioration of PS and never initiated treatment, leaving 32 patients who received at least 1 dose of protocol therapy (Table 2). A total of 154 cycles of therapy were administered (median of 6 cycles per patient); 19 patients completed all 6 planned cycles of treatment. Reasons for not completing all 6 cycles of therapy were objective disease progression (n = 6), symptomatic disease progression (n = 3), adverse events (n = 3), and patient refusal (n = 2).

Table 2. Treatment Received

No. of Cycles Administered	No. of Patients (n=32)
1	1
2	4
3	3
4	3
5	2
6	19
Gemcitabine dose intensity	Planned: 666.7 mg/m^2/wk
	Actual (median): 531.3 mg/m^2/wk
Cisplatin dose intensity	Planned: 23.33 mg/m^2/wk
	Actual (median): 21.01 mg/m^2/wk
Carboplatin dose intensity	Actual (median): 130.1 mg/wk

Five patients initiated therapy with carboplatin (2 due to decreased creatinine clearance, 1 due to baseline hearing deficit, and 2 due to physician choice). An additional 5 patients were switched as per protocol from cisplatin to carboplatin during therapy (2 due to decreased creatinine clearance [Cycles 2 and 4]; 1 due ototoxicity [Cycle 5]; 1 due to protracted cisplatin-induced vomiting [Cycle 1]; and 1 after a seizure, believed possibly related to cisplatin, in Cycle 1). Approximately one-third and one-half of patients received >90% of planned dose intensity for gemcitabine and cisplatin, respectively. The vast majority of dose modifications and delays were due to hematologic toxicity.

Toxicity and Adverse Events

Reported adverse events considered by the investigator to be related to protocol therapy are summarized in Table 3; hematologic toxicity is summarized in Table 4. Observed toxicity was within the range of what would be expected from the combination of cisplatin and gemcitabine. Overall, the therapy was well-tolerated, with only 3 patients withdrawing from therapy due to adverse events (1 patient each with grade 4 dyspnea/grade 2 pneumonitis, grade 3 thrombocytopenia, and grade 3 fatigue). There were no episodes of febrile neutropenia reported. There was no clinically relevant biochemical toxicity, and there were no deaths attributable to study therapy.

Table 3. Reported, Related Nonhematologic Adverse Events Occurring in at Least 10% of Patients or of at Least Grade 3 in Severitya

	No. of Patients With Gradeb
	1	2	3	4
a Adverse event considered by the investigator to be at least “possibly” related to protocol therapy. b Toxicity was assessed using the National Cancer Institute Common Toxicity Criteria (version 2.0).
Auditory
Inner ear/hearing	4	4	—	—
Cardiovascular
Thrombosis/embolism	—	—	3	2
Constitutional
Fatigue	8	11	2	—
Gastrointestinal
Anorexia	1	3	—	—
Constipation	4	—	—	—
Dysguesia	4	1	—	—
Mucositis	3	2	—	—
Nausea	16	7	1	—
Vomiting	6	4	1	—
Hemorrhage
Other hemorrhage	—	—	1	—
Neurologic
Neuropathy, motor	1	1	—	—
Neuropathy, sensory	6	—	—	—
Seizure	—	1	—	—
Pain
Headache	2	2	1	—
Pulmonary
Dyspnea	—	3	—	1
Other
Alopecia	6	2	—	—

Table 4. Hematologic Toxicity

	Worst, Any Cycle	Worst CTC Grade (No. of Patients)a
	Median Nadir (Range)	0-2	3	4
CTC indicates Common Toxicity Criteria. a Toxicity was assessed using the National Cancer Institute Common Toxicity Criteria (version 2.0).
Leukocytes (×10^9/L)	1.8 (0.6-5.5)	15	16	1
Granulocytes (×10^9/L)	0.90 (0.1-4.4)	13	14	5
Hemoglobin (g/L)	91 (68-133)	27	5	0
Platelets (×10^9/L)	62 (12-196)	21	11	0

Antitumor Efficacy

Of the 33 eligible patients, 3 were not evaluable for response: 1 patient who never received protocol therapy and 2 in whom the appropriate radiologic investigations were not performed. Although the protocol called for the enrollment of 34 response-evaluable patients, an insufficient number of responses had been observed after the assessment of 30 such patients to warrant continuation, because the 13 responses required to declare the regimen active could not be met. One complete response (CR) and 7 partial responses (PR) were observed, for an objective response rate of 24% (95% confidence interval, 11-42%) among the 33 eligible patients. The duration of response ranged from 1.3 to 11. 3 months, with a median of 6.7 months. Responses were observed in 3 patients with adenocarcinoma (including the CR); 2 with ACC; and 1 each with acinic cell carcinoma, MEC, and salivary duct carcinoma. In the 5 patients who initiated therapy with carboplatin, no objective responses were observed.

Three patients were considered to have objective tumor progression by RECIST at the first response assessment, whereas 19 patients had stabilization of disease as their best response, for a median duration of 6.7 months (range, 1.8-16.2 months). Of these 19 patients, 12 with a variety of histologic subtypes demonstrated some degree of tumor shrinkage insufficient to meet RECIST criteria for a PR (Fig. 1, waterfall plot: target/measurable lesions only).

Figure 1. Antitumour efficacy is shown. An asterisk indicates those patients with an objective response as per Response Evaluation Criteria In Solid Tumors (RECIST).

The median survival for all 33 eligible patients was 13.8 months, with a 1-year overall survival rate of 57%.

DISCUSSION

Cisplatin-based combination chemotherapy is commonly used as palliative therapy for patients with advanced salivary gland cancers. The greatest reported experience is with the regimen of cisplatin, doxorubicin, and cyclophosphamide (CAP), which leads to objective responses in ≤33% of patients with ACC.12 Other cisplatin-based regimens have likewise led to objective responses, although to the best of our knowledge only single studies exist for these other combinations. Although it may slightly increase the chance of an objective tumor response, it is not fully clear that combination therapy offers a therapeutic advantage over therapy with single agents that have documented activity, such as mitoxantrone,13, 14 epirubicin,15 paclitaxel,16 and vinorelbine.17 Thus, the choice of a combination regimen, particularly 1 containing cisplatin, must bear in mind the likelihood of additional toxicities.

As a single-agent, cisplatin has to our knowledge been evaluated prospectively in only 1 clinical trial18; 2 objective responses were observed in 13 patients with ACC, 1 in 5 patients with MEC, and none in 5 with adenocarcinoma. At the time of initiation of this trial, there were no published data regarding the role of gemcitabine in salivary gland cancers. A recent trial of single-agent gemcitabine in 21 patients with advanced ACC revealed no objective tumor responses, with 8 of the patients found to have disease progression at the first response assessment.19 A very brief mention of an ongoing trial of gemcitabine in advanced salivary gland cancers indicated there were no objective responses in 16 patients who had been enrolled at the time of last follow–up, including 8 with ACC.20 Thus, the data do not support the use of single-agent gemcitabine in those patients with ACC, and to our knowledge there are no available data on its role in salivary cancers of other histologies.

This study revealed an objective response rate of 24% for the combination and did not meet the predefined criteria for being declared an active regimen. However, an additional 12 patients, although not meeting the RECIST criteria for a PR (and who were therefore coded as having stable disease [SD]), had some degree of tumor shrinkage. In malignancies that may have a protracted, indolent course, stabilization of disease may not necessarily reflect efficacy of the agent under study, particularly if documented progressive disease is not a criterion for study enrollment; however, the observation of any degree of tumor shrinkage likely does reflect some degree of antitumor activity. Direct comparisons of the relative efficacy of chemotherapies for advanced salivary gland cancer are difficult, due to the small number of trials, the relatively few patients enrolled in each trial (particularly with non-ACC histology), and differences in eligibility criteria, such as the requirement for documented disease progression before enrollment. Many trials do not report the outcomes separately for each histologic subtype, and none, to our knowledge, have reported what proportion of patients with SD demonstrated any degree of response.

Responses were observed in all common histologic subtypes, and the proportion of those with ACC responding (2 of 10) was comparable to that noted with single-agent cisplatin (2 of 13). Given the lack of single-agent activity of gemcitabine, it does not appear that this combination offers an advantage over other cisplatin-based regimens (particularly CAP), or indeed single-agent cisplatin, for patients with ACC histology. In contrast, the regimen may have promising activity in patients with adenocarcinoma histology, with 3 confirmed objective responses and 3 additional patients demonstrating tumor shrinkage; this differential responsiveness of the different histologic subtypes to various chemotherapeutic agents has been described by others.16 There are too few patients to draw any conclusions regarding the use of this regimen in patients with the other histologic subtypes.

To the best of our knowledge, there are no data available regarding the activity of single-agent carboplatin in patients with advanced salivary gland cancers. In what to our knowledge is the sole trial of a carboplatin-containing regimen, only 2 of 10 patients with ACC responded to the combination of carboplatin and paclitaxel.21 Given the lack of responses observed in the 5 patients in the current trial who initiated therapy with carboplatin, the available data do not support the routine substitution of carboplatin for cisplatin in the treatment of advanced salivary gland cancers.

Overall, the regimen was well-tolerated, with an adverse event profile that was consistent with previous experience with this combination. Although currently this regimen is usually administered with cisplatin on Day 1 and gemcitabine on Days 1 and 8 of a 21-day cycle, at the time this trial was designed, there was some controversy as to the optimal dose and schedule. It is now clear that, in general, a 21-day cycle permits a higher dose intensity of gemcitabine than a 28-day schedule, in which gemcitabine is administered on Days 1, 8, and 15.22 Preclinical studies vary with regard to whether the additive cytotoxicity of cisplatin and gemcitabine is augmented by either concomitant administration or spacing the agents by 24 hours. In a 28-day schedule, there is some evidence to suggest that administering cisplatin on either Day 2 or 15 leads to improved efficacy in patients with advanced nonsmall cell lung cancer.23 However, in the absence of convincing data of a benefit to spacing the agents in a 21-day schedule, the administration of both drugs on Day 1 would increase the convenience and acceptability of the regimen.

In the current study, the combination of cisplatin and gemcitabine demonstrated some modest activity in patients with advanced salivary gland cancers, but did not meet the predefined criteria to warrant further study, although this regimen may be of interest in patients with the adenocarcinoma subtype. Because there is no standard systemic therapy for advanced salivary gland cancers, the enrollment of these patients into clinical trials remains a high priority. Given their relative chemoresistance, there is a need to better understand the molecular pathways of relevance in salivary gland cancers, so as to be able to incorporate targeted agents with, arguably, a better therapeutic index into the palliative management these incurable malignancies.

CONFLICT OF INTEREST DISCLOSURES

The NCIC Clinical Trials Group (CTG) is supported by a grant from the Canadian Cancer Society. Eli Lilly Canada Inc (Toronto, Ontario, Canada), kindly provided gemcitabine (but no financial support) for this trial.