SEARCH

SEARCH BY CITATION

Keywords:

  • head and neck cancer;
  • chemotherapy;
  • radiation;
  • registry

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

BACKGROUND:

A study was undertaken to examine the patterns of systemic therapy use in conjunction with radiation therapy for patients with locally advanced head and neck squamous cell cancer.

METHODS:

Between December 1, 2005 and May 11, 2009, 2874 patients with newly diagnosed head and neck squamous cell cancer who were scheduled to receive radiotherapy and/or drug therapy were registered in a prospective national database. The database was specifically analyzed to examine patients who received chemotherapy in conjunction with definitive radiotherapy.

RESULTS:

A total of 1144 patients received systemic therapy in conjunction with radiotherapy; 645 (56%) patients received agents concurrent with radiation therapy, 49 (4%) patients received chemotherapy before radiotherapy (induction), 224 (20%) patients received chemotherapy before and during radiotherapy (sequential), and 226 (20%) patients received chemoradiation after surgery. Single-agent cisplatin, single-agent cetuximab, and carboplatin plus paclitaxel were, in order, the 3 most commonly prescribed concurrent regimens. Concurrent cisplatin was more frequently used in the academic setting compared with the community setting (P = .0015). Postoperative chemoradiation, rather than radiation alone, was more commonly used in academic centers compared with community practice centers (P = <.0001).

CONCLUSIONS:

The LORHAN (Longitudinal Oncology Registry of Head and Neck Carcinoma) database is a useful barometer of current US practice patterns and can be applied to analyze future trends in the combined modality management of head and neck cancer. Sequential chemotherapy is used frequently, but cisplatin-based concurrent chemoradiation remains the most commonly used regimen for locally advanced head and neck cancer. Cancer 2011. © 2010 American Cancer Society.

At the time of diagnosis, the majority of patients with head and neck squamous cell carcinoma (HNSCC) present with locoregionally advanced disease in which the cure rate with surgery or radiation alone is achievable but many would consider inadequate. Combined modality treatment with systemic agents and radiation is a potential treatment option for these patients. Randomized phase 3 studies have demonstrated the superiority of radiation therapy plus concurrent cisplatin (100 mg/m2 on Days 1, 22, and 43) compared with radiation alone in the treatment of unresectable locally advanced HNSCC and for organ preservation or postoperative adjuvant therapy in the presence of high-risk pathologic features, particularly positive margins or extranodal extension.1-4 The use of radiation therapy plus concurrent antiepidermal growth factor receptor monoclonal antibody, cetuximab, has also demonstrated improved outcomes for patients with locally advanced HNSCC compared with radiation alone.5 Emerging data with induction and sequential chemotherapy regimens have shown promising results, but have also served to add to the confusion regarding the optimal treatment approach for patients with locally advanced HNSCC.6, 7 Availability of new radiation techniques and altered fractionation schedules offer further advancements for patients, but can further complicate decision making regarding selection of systemic therapy.8-10

In the routine management of locally advanced HNSCC, it is uncertain to which degree oncologists adhere to evidenced-based practices such as defined by randomized controlled phase 3 studies (level 1 evidence) or by category 1-based National Comprehensive Cancer Network guidelines. No large-scale databases are available that have accurately tracked information regarding chemotherapy utilization in conjunction with radiation for locally advanced HNSCC. Such data are pertinent to the design of future clinical trials, but are also critical to gain understanding as to the impact of the results of important clinical trials upon the routine management of HNSCC.

LORHAN (Longitudinal Oncology Registry of Head and Neck) is a prospective national registry designed to examine patterns of care of head and neck cancer patients in the United States. The LORHAN database collects baseline disease-specific information and treatment data including chemotherapy and radiation therapy, from community as well as academic practices. We conducted the following analysis of the LORHAN database to examine current practice patterns of US physicians who prescribed chemotherapy in conjunction with radiation therapy for newly diagnosed HNSCC.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Study Population

The LORHAN database was started in 2005. Patients eligible for enrollment into the registry must have had a new diagnosis of cytologically or histologically confirmed carcinoma of the oral cavity, oropharynx, hypopharynx, nasopharynx, larynx, or neck lymph node metastasis from an occult primary tumor. Patients must have been a candidate for chemotherapy and/or radiation therapy at the time of study registration, be aged ≥18 years, and have signed an institutional review board-approved consent. Patients participating in a therapeutic clinical trial were eligible for entry in the LORHAN registry.

Data entry in the LORHAN registry was electronic, incorporating encrypted point-to-point data transfer via secure HTTP protocols. For patients who received chemotherapy, data were collected regarding type of chemotherapy, dose, number of cycles, and timing of chemotherapy with respect to radiation and surgery. Other core baseline data collected for the registry included demographics, clinical characteristics, radiation treatment. Data forms were collected at predesignated time points: at baseline, after completion of initial treatment, 1 year, 2 years, 3 years, and/or at the time of disease progression, relapse, or development of a new primary cancer. Entry of data from each time point was locked only after reviewed for accuracy by the investigator at the institutional site. Compensation to centers was provided on a per case basis to cover regulatory fees, personnel, and data management costs. Payment was dispensed in increments after locking of data at predesignated time points.

Participating institutions were self-designated as being community practice or academic centers. Patients from 103 participating institutions (22 academic, 81 community), with broad representation of all geographic regions of the United States, were enrolled in the LORHAN registry between December 1, 2005 and May 11, 2009. Patients were included in this analysis if they had completed and locked baseline and initial therapy forms. Patients included in the present analysis must also have received systemic therapy (including classic cytotoxic drugs or biologic targeted agents) in conjunction with definitive radiation therapy.

Statistical Analysis

Statistical analyses used chi-square, t test, and other nonparametric tests as appropriate, for 2 sample comparisons. All tests were performed under a 2-sided null hypothesis with a 2-tailed P value of ≤.05 to reject the null. All analyses were performed using R (The R Foundation for Statistical Computing), version 2.11.0 or greater.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Patient Characteristics

At the time of analysis, 2874 patients were registered in the LORHAN database, of whom 2363 patients had locked baseline assessments and 1747 patients had completed and locked initial therapy forms. Characteristics of patients with locked baseline assessments are shown in Table 1 for all patients and subdivided by type of institution in which patients were treated—academic or community-based. Participating institutions have enrolled patients from 35 states, representing all major regions of the United States. More patients (n = 1009) were registered from academic institutions than from community practice sites (n = 738).

Table 1. Baseline Patient Characteristics
CharacteristicTotal (%)Academic (%)Community (%)
  1. PS indicates performance status.

No.1747 (100%)1009 (100%)738 (100%)
Age, mean y59.257.461.8
Sex
 Women412 (24%)228 (23%)184 (25%)
 Men1335 (76%)781 (87%)554 (75%)
Race
 White1443 (83%)817 (81%)626 (80%)
 Black215 (12%)137 (14%)78 (12%)
 Asian19 (1%)9 (0.8%)10 (1.3%)
 Native American, Alaskan, Hawaiian, Pacific Islander17 (1%)12 (1%)5 (0.7%)
 Unknown/other10 (0.5%)9 (0.8%)1 (0.1%)
Zubrod PS
 0694 (37%)361 (36%)333 (45%)
 1732 (42%)446 (44%)286 (39%)
 2266 (15%)170 (17%)96 (13%)
 ≥355 (3%)32 (3%)23 (3%)
Disease site
 Oropharynx697 (40%)431 (43%)266 (36%)
 Larynx383 (26%)176 (17%)207 (28%)
 Oral cavity282 (16%)170 (17%)112 (15%)
 Unknown primary89 (5%)55 (5%)34 (5%)
 Salivary77 (4%)50 (5%)27 (4%)
 Nasopharynx73 (4%)34 (3%)39 (5%)
 Hypopharynx70 (4%)40 (4%)30 (0.4%)
 Paranasal sinus36 (2%)29 (3%)7 (1%)
 Nasal15 (1%)11 (1.5%)4 (0.5%)
 Other25 (1%)13 (1.2%)12 (3%)
Stage
 I117 (7%)42 (4%)75 (10%)
 II135 (8%)73 (7%)62 (8%)
 III275 (16%)136 (16%)139 (19%)
 IV1048 (60%)674 (67%)374 (51%)
 Distant metastasis34 (2%)18 (1.7%)16 (2%)

Distribution of Treatment Modalities

A total of 1617 (93%) patients received radiation as a component of their therapy regimen; 468 (27%) patients received radiation therapy alone or surgery followed by radiation therapy. Of the 1144 (65%) patients who received chemotherapy in conjunction with radiotherapy, 645 (56%) patients received chemotherapy concurrent with radiation therapy, 49 (4%) patients received chemotherapy before radiotherapy (induction), 224 (20%) patients received chemotherapy before and during radiotherapy (sequential), and 226 (20%) patients received chemoradiation after surgery (see Table 2). To examine differences in the pattern of utilization of these treatment modalities between academic and community practices, relative uses of each modality within each group were compared. No differences were observed in the relative frequency of use of concurrent chemoradiation or sequential chemotherapy. Overall, induction chemoradiation was not frequently used but was disproportionately more commonly used in the community setting (P = .046).

Table 2. Initial Treatment Analysis
Treatment ModalityTotal, n = 1144, No. (%)Intergroup Comparison
  Academic, n = 671, No. (%)Community, n = 473, No. (%)P
  1. CRT indicates concurrent chemotherapy plus radiation therapy; IC, induction chemotherapy; SQ, sequential chemotherapy; Post-Op CRT, surgery followed by concurrent chemotherapy plus radiation therapy.

CRT645 (56%)360 (54%)285 (60%).027
IC49 (4%)22 (3%)27 (6%).046
SQ224 (20%)129 (19%)95 (20%).718
Post-Op CRT256 (22%)160 (24%)66 (14%)<.0001

Chemotherapy Regimens

Table 3 shows initial prescribed regimens for patients who received systemic treatment in conjunction with radiation. For patients who received radiation therapy with concurrent systemic therapy, cisplatin alone was the most commonly prescribed regimen and was more commonly used in the academic setting compared with the community setting (P = .0015). Of patients who received single-agent cisplatin, approximately 15% received a weekly dosing schedule of cisplatin (data not shown). The next most common regimens were single-agent cetuximab, followed by carboplatin plus paclitaxel; no intergroup differences were observed with respect to the use of these regimens between institution type.

Table 3. Most Common Chemotherapy Regimens
RegimenTotal, n = 1144, No. (%)Intergroup Comparison
  Academic, n = 671, No. (%)Community, n = 473, No. (%)P
  1. 5-FU indicates 5-fluorouracil.

Cisplatin584 (51%)369 (55%)215 (16%).0015
Cetuximab243 (21%)135 (20%)108 (23%).2691
Carboplatin + paclitaxel83 (7%)48 (7%)35 (7%).8745
Cisplatin + docetaxel + 5-FU54 (5%)19 (3%)35 (7%).0003
Cetuximab+ cisplatin + docetaxel + 5-FU33 (3%)33 (5%)0 (0%)<.0001
Carboplatin30 (3%)17 (3%)13 (3%).8228
Cisplatin + 5-FU17 (2%)2 (0%)15 (3%)<.0001
Cisplatin + paclitaxel14 (1%)6 (1%)8 (2%).23
Other86 (8%)42 (6%)44 (9%).0546

When chemotherapy was used before radiation therapy (for induction or sequential treatment) the most common regimens, in descending order, were: cisplatin, 5-fluorouracil, plus docetaxel; cisplatin, 5-fluorouracil, docetaxel, plus cetuximab; and cisplatin, 5-fluorouracil, plus paclitaxel. Sequential chemotherapy was frequently used in both academic and community settings, as shown in Table 2.

To examine whether specific patient characteristics influence the selection of concurrent chemotherapy, we performed a logistic regression analysis of the most commonly used systemic therapy regimens (cisplatin and cetuximab), as shown in Table 4. Use of cetuximab, relative to cisplatin, was more common in patients with increasing age, and Zubrod performance status (PS) ≥2. Cisplatin, relative to cetuximab, was more commonly prescribed in patients with nasopharyngeal cancer.

Table 4. Logistic Regression Analysis: Characteristics Associated With Use of Single-Agent Cetuximab Versus Cisplatin
CharacteristicEstimateStandard ErrorOdds Ratio95% CIzP
  1. CI indicates confidence interval.

Community setting−4.920.601.160.83-1.630.89.37
Male sex−0.0030.170.990.67-1.50−0.02.99
Age0.0080.0081.061.04-1.086.99<.0001
Zubrod 10.210.191.240.86-1.781.14.25
Zubrod 20.770.242.161.33-3.493.14.002
Zubrod ≥31.270.553.571.19-10.862.30.02
Larynx−0.040.310.960.53-1.77−0.12.90
Nasopharynx−2.841.060.060.003-0.33−2.68.007
Oral cavity0.570.301.770.98-3.251.87.06
Oropharynx0.340.261.400.84-2.381.27.20
Enrolled in clinical trial−0.070.210.920.61-1.38−0.38.71

To examine factors associated with tolerability of concurrent systemic therapy, a logistic regression analysis of characteristics associated with dose reduction of either cisplatin or cetuximab was performed. Overall, no difference in the rate of dose reductions was observed between single-agent cisplatin and single-agent cetuximab (data not shown). As shown in Table 5, for patients receiving either cisplatin or cetuximab, chemotherapy dose reductions were less frequent in patients treated in the community setting, and in patients who were treated on a clinical trial. In patients with Zubrod PS ≥3, systemic therapy dose reductions were significantly more common.

Table 5. Logistic Regression Analysis: Characteristics Associated With Dose Reduction in Either Single-Agent Cetuximab or Cisplatin
CharacteristicEstimateStandard ErrorOdds Ratio95% CIzP
  1. CI indicates confidence interval.

Community setting−0.640.170.530.37-0.7452−3.60.0003
Male sex−0.020.200.980.66-1.45−0.11.91
Age−0.00020.0080.0080.98-1.02−0.03.98
Zubrod 1−0.0090.180.990.69-1.410.05.96
Zubrod 20.220.241.240.77-2.010.90.37
Zubrod ≥31.230.563.521.17-11.12.23.02
Larynx−0.060.300.940.52-1.70−0.20.84
Nasopharynx0.470.391.600.73-3.491.19.24
Oral cavity0.160.301.180.65-2.140.53.59
Oropharynx0.140.251.150.71-1.910.56.57
Enrolled in clinical trial−0.870.230.420.26-0.65−3.73.0002

Postoperative Adjuvant Therapy

Table 6, presents the distribution of treatment by center type, among patients who received postoperative adjuvant therapy (n = 364). The use of chemoradiation was higher in the academic setting at 66% as compared with 56% in the community setting, Correspondingly, the use of radiation alone was higher in the community setting at 34% as compared with 46% in the academic setting (Table 6; P value = .0179). We examined systemic agents used concurrently with radiation after surgery. As shown in Table 7, only single-agent cisplatin or cetuximab was commonly used in concurrent chemotherapy regimens. In comparison, cisplatin was more commonly used at academic centers than in community centers, whereas, with respect to cetuximab use, no difference was observed based on center type.

Table 6. Postoperative Radiation Versus Postoperative Chemoradiation
SettingPostoperative Therapy
Radiation AloneChemoradiationTotal = 364
Academic Institutions81 (34%)160 (66%)241
Community Practice Institutions57 (46%)66 (56%)123
Table 7. Most Common Postoperative Concurrent Drug Regimens
RegimenTotal, n = 226Intergroup Comparisons
Academic, n = 160Community, n = 66P
  1. CI indicates confidence interval.

Cisplatin153/226115/16038/66.0366, χ2 (1) = 4.37
% [CI]67.7 [61.4-73.5]71.9 [64.5-78.3]57.6 [45.6-68.8] 
Cetuximab48/22631/16017/66.2861, χ2 (1) = 1.14
% [CI]21.2 [16.4-27.0]19.4 [14.0-26.2]25.8 [16.7-37.4] 
Other25/22614/16011/66.0845, χ2 (1) = 2.98
% [CI]11.1 [7.6-15.8]8.8 [5.3-14.2]16.7 [9.6-27.4] 

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

In the present study we used a nationwide, prospective, head and neck cancer database to analyze the current practice patterns of US oncologists with respect to utilization of systemic therapy in conjunction with radiation therapy for head and neck cancer. Unlike other available databases, the LORHAN registry collects patient data prospectively from the time of the initiation of definitive therapy. In contrast, Surveillance, Epidemiology and End Results (SEER) and National Cancer Database do not track detailed information regarding radiation, chemotherapy drugs, supportive care, and outcomes. Hospital and pharmacy data can be used to examine utilization of chemotherapy, but not in a disease-specific manner or in a way such that chemotherapy use, delivered concurrent with radiation, can be captured. Likewise, the data from SEER and National Cancer Database are limited by lacking association of specific treatment data with detailed disease-specific clinical information. Increasing availability of the electronic medical documentation with integrated chemotherapy order entry software holds the promise of disease-specific data extraction and outcomes analysis. However, retrospective data extraction from the electronic medical record is severely limited because of the incompleteness of data. Furthermore, wide scale multi-institutional analysis is hampered by the use of multiple software platforms that further magnify the problem of inhomogeneous data collection. The prospective design and careful selection of predesignated data fields of the LORHAN database overcome these pitfalls and permit the usefulness of the data.

Identification of patient subgroups that are more likely to achieve maximal benefit from concurrent chemoradiation versus from induction/sequential chemoradiation has not been well defined in the literature. Considerable controversy presently exists regarding the relative benefits of concurrent cisplatin-based chemoradiation versus induction or sequential therapy.11 Two large phase 3 clinical trials using induction and sequential have recently been reported.6, 7 However, these trials only compared the effectiveness of specific induction chemotherapy regimens, and did not examine the inherent value of induction chemotherapy versus no induction. The existing data from large phase 3 studies supports the notion that concurrent cisplatin-based chemoradiation is superior to radiation alone and represents the highest level of evidence for management of locally advanced HNSCC.1-4 It has been speculated that the use of induction chemotherapy has proliferated in the routine management of locally advanced HNSCC, despite lack of randomized, controlled phase 3 studies (level I evidence) directly comparing induction (or sequential) therapy versus concurrent chemoradiation. Our analysis confirms this impression. Furthermore, because the LORHAN registry is ongoing, we will be able to compare future trends with past patterns of induction/sequential therapy use such that, for instance, the impact of future phase 3 studies can be assessed.

Level I evidence supports the use of concurrent cisplatin as well as concurrent cetuximab; we show also the frequent use of other concurrent chemotherapy regimens, for which only lower level of supporting evidence exists. The vast majority of patients (80%) in the registry had optimal PS with Zubrod scores ≤1. However, approximately 20% of registered patients had suboptimal PS (Zubrod ≥2), which in part may explain the use of potentially less intensive, and less well studied, concurrent regimens. Forty-eight percent of LORHAN registry patients (excluding nasopharynx, salivary, unknown primary, and sinus cancer cases) had oropharyngeal carcinoma. Based upon studies that have examined human papillomavirus (HPV) in squamous cell carcinoma, showing that approximately 60% of oropharyngeal carcinomas are associated with HPV, we can estimate that as much as 30% of registry patients may have HPV-related disease.12 We speculate that physician preference for using noncisplatin regimens may result as an attempt to tailor less toxic treatment to this more favorable prognostic group. The LORHAN registry tracks data as to whether HPV testing is performed. It will therefore be possible to test this hypothesis as the use of HPV testing becomes more widespread.

Two phase 3 studies have demonstrated the superiority of postoperative concurrent cisplatin plus radiation therapy compared with postoperative radiation alone.3, 4 Combined analysis of these trials, EORTC 22931 and RTOG-Intergroup trial 9501, identified poor prognostic factors that enable prediction of patients most likely to benefit from postoperative chemoradiation.13 We observed that in academic centers, compared with community centers, postoperative chemoradiation was more frequently used than postoperative radiation alone. Definitive conclusions regarding this observation are limited because of the lack of other information (such as extracapsular lymph node extension or surgical tumor margin status) that would allow further analysis. However, we can speculate that this observation may be because of more high-risk patients referred to academic centers for surgical therapy. Alternatively, it is possible that high-risk patients are proportionately equally represented in the community setting, but that the pattern of practice is divergent from evidence-based guidelines such as category 1 National Comprehensive Cancer Network guidelines.14 Finally, the gain achieved with chemoradiation over radiation alone in the postoperative setting is modest, and comes at the cost of increased toxicity profiles. It may be that concerns about toxicity, as well as the expertise of centers to nurture patients through challenging treatment regimens, influence the use of combined regimens at various centers.

Although the information presented here provides a representation of the current pattern of utilization of chemotherapy for the multimodality management of head and neck cancer in the United States, we recognize the limitations of these data. Academic and community practice centers that voluntarily agreed to participate in the LORHAN registry may have done so as a result of a particular interest or expertise in the field of head and neck oncology. Therefore, it is conceivable that the database may reflect biases in practice patterns that may not necessarily be representative of the general US oncology community. It is also possible that some participating centers were selective in the decision to register qualifying patients; however, it is likely that this potential bias is diluted given the large number of patients and participating centers in the registry. An ideal head and neck cancer registry would include all HNSCC patients rather than selected patients who meet eligibility criteria; this is however not practical. For the purpose of this study, it is not necessary for each participating center to enroll every patient to gauge their pattern of practice. Compared with SEER or National Cancer Database, it is plausible that the LORHAN registry may not accurately reflect the demographic features of the US head and neck cancer patient population. However, detailed demographic information was captured in the LORHAN registry and appears to be consistent with other head and neck cancer demographic reports. Comprehensive analysis of demographic data and association with clinical outcome will be the subject of future inquiry and is beyond this scope of this report.

In summary, we report, to our knowledge, the first broad snapshot of the practice patterns of US oncologists regarding the utilization of systemic therapy in conjunction with radiation and/or surgery for locally advanced HNSCC. We show that cisplatin-based concurrent chemoradiation is the most frequently used regimen, and that concurrent cetuximab plus radiation is the next most commonly used regimen; both regimens are consistent with category 1 level of evidence in National Comprehensive Cancer Network guidelines.14 A substantial proportion of patients in the United States receive induction or sequential chemoradiation for locally advanced HNSCC despite a lower level of evidence to support their use. We conclude that the LORHAN database is a useful tool to gauge current practice patterns and will be valuable in the future for comparison of evolving trends of the combined modality management of head and neck cancer. The data presented here may contribute to the design of future clinical trials. Moreover, the impact of major clinical trials may also be assessed using this prospective database.

CONFLICT OF INTEREST DISCLOSURES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Supported by Eli Lilly and Company, Indianapolis, Indiana (formerly by Imclone Systems Inc., Branchburg, NJ). Financial disclosures: Stuart Wong: none. Paul Harari: consultant/advisory board, Amgen, Astra Zeneca, Imclone, Genentech, Bristol Myers Squibb. Adam Garden: advisory board, Imclone. Marc Schwartz: none. Lisa Bellm: none. Amy Chen: Med Net. Walter Curran: Imclone. Barbara Murphy: advisory/consultant, Bristol Myers Squibb. Kian Ang: Imclone.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES