Concurrent Chemoradiation for Cancer of the Cervix: Results of a Multi-Institutional Study From the Setting of a Developing Country (India)

Purpose The primary output of hospital-based cancer registries is data on cancer stage and treatment-based survival that can be used to evaluate patient care, but because there are many challenges in obtaining follow-up details, a separate study on patterns of care and patterns of survival for patients at selected sites was initiated under the National Cancer Registry Programme of India. This article presents the results for cervical cancer. Patients and Methods A standardized patient information form was used to record patient information, and data were entered into a central repository—the National Centre for Disease Informatics and Research. The study patients were from 12 institutions and were diagnosed between January 1, 2006, and December 31, 2008. Patterns of treatment were assessed for 7,336 patients, and patterns of survival were determined for 2,669 patients from six institutions, at least 70% of whom had data regarding follow-up as of December 31, 2012. Results Of 7,336 patients, 55.5% received optimal radiotherapy (RT). In all, 80.9% of patients had locally advanced cancers (stage IIB to IVA), 51.1% received RT alone, and 44.4% received concurrent chemoradiation (RTCT). In 1,753 patients with locally advanced cancers, significantly better survival was observed with RTCT than with RT alone (5-year cumulative survival, 70.2% v 47.3%; hazard ratio, 0.48; 95% CI, 0.41 to 0.56). Conclusion A conservative estimate indicates that, on an annual basis, 38,771 patients with cervical cancers in India alone do not get the benefit of RTCT and thus they have poorer survival. There is a need to reiterate the National Cancer Institute's alert that advised supplementing chemotherapy to radiation for locally advanced cancer of the cervix in the context of the developing world, where 84.3% of cancers of the cervix occur.


INTRODUCTION
Cervical cancer comprises 7.92% of cancers in women worldwide, 1 and in India alone, estimates indicate that there will be approximately 94,000 new cases per year. 2 Information on care for patients with cancer and patient survival is essential in assessing cancer treatment services, and a hospital-based cancer registry (HBCR) is central to this effort. 3 In developing countries, follow-up after treatment presents many challenges. 4 The main aim of this study on patterns of cancer care and survival (POCCS) was to obtain information on treatment based on clinical stage and on survival for patients with cancers of the cervix, breast, and head and neck. This study of POCCS presents results on cervical cancer.
An enhanced version of the prior technique 15 for capturing electronic data and using the Internet to transmit that data to a central repository constituted the basic design and framework on which the required clinical information was obtained.

PATIENTS AND METHODS
Twelve institutions (centers) participated in the study. The names of the institutions along with members of the Patterns of Cancer Care and original report abstract Survival Group are provided in the Appendix (online only). A standardized patient information form (Data Supplement) created by oncologists with specific expertise in treating cancer of the cervix was hosted on the Hospital-Based Cancer Registries Web site. A printed form with instructions was supplied for each study participant. Trained staff completed the form by using patient and/or attendant interviews, by scrutinizing medical records and other relevant documents and registers, and by having discussions with concerned clinicians. Collaborating centers were given individual login IDs and passwords along with instructions for entering data online to be electronically transmitted to a central repository-the National Centre for Disease Informatics and Research (NCDIR). The mandate and mission statements of this one-of-akind center (an outcome of the National Cancer Registry Programme of the Indian Council of Medical Research) are provided at the NCDIR Web site.

Selection Criteria
Treatment patterns based on cancer stage were examined for 7,336 newly diagnosed (January 1, 2006, to December 31, 2008) patients with cervical cancer treated at their respective institutions. However, survival analysis was restricted to data from six centers that had follow-up information for at least 70% of their respective patients as of December 31, 2012. The total number of patients was 2,686, but 17 had no details on follow-up after the date of last treatment; therefore, they were excluded, leaving 2,669 patients. Some institutions had details on follow-up for more than 70% of their patients; thus, the overall pooled percentage of patients with follow-up information for survival analysis was 87%.
The main end point, overall survival, was defined as date of diagnosis to date of death from any cause (when death was before January 1, 2013). All other patients were regarded as alive, and the last date of follow-up was the censored date. The number and proportion of patients with toxicity (in both early and late complications) and recurrence are based on any one such reported event.
Only squamous cell carcinomas are included. All centers followed the International Federation of Gynecology and Obstetrics (FIGO) staging system. 16 Detailed survival analysis and discussion are focused on locally advanced cervical cancer (FIGO stage IIB to IVA) because advances in treatment (especially concurrent chemoradiation [RTCT]) relate to this category, which involved 72.3% of the patients (ie, 1,930 of 2,669).

Radiotherapy
The standard practice of administering approximately 50 Gy (total dose in 20 to 25 fractions) radiation to the entire pelvis was followed. 17,18 Other parameters such as the use of intracavity brachytherapy, use of a radiotherapy (RT) machine (linear accelerator or cobalt-60), number of fields, and duration/fractionation of RT were also considered. To simplify analysis, the term "optimal radiotherapy," as outlined by the Chemotherapy for Cervical Cancer Meta-Analysis Collaboration 19 and Shrivastava et al 18 was used for this study. Optimal RT is defined as administering at least 45 Gy by external beam (minimum of 20 fractions) plus intracavity brachytherapy (any dose). All other types of RT were classified as suboptimal. The RT machine used and beam arrangement (fields) were separately factored and were adjusted for in the statistical analysis. Only a few patients received RT via intensity-modulated RT, image-guided RT, or other types of RT, and these factors were disregarded.

Chemotherapy
Chemotherapy (CT) was administered within 1 week before or 1 week after the start of RT in 89.1% of the patients. The predominant protocol was monotherapy with cisplatin (cis-dichlorodiammine platinum). Patients who received other drugs alone or in combination with cisplatin were grouped separately. The total dose of cisplatin, the number of cycles, and dose in mg/m 2 per cycle were calculated. However, to simplify analysis, we used only the total dose of cisplatin given. Dose was grouped a priori into less than 150 mg, 150 to 199 mg, 200 to 239 mg, and Ն 240 mg. The mean and median number of cycles was four, and 66.2% of patients received four or more cycles. The average computed dose per cycle was 40 mg/m 2 per week. More than 75% of patients received at least 39 mg/m 2 cisplatin per week. The median total dose of cisplatin was 200 mg, and the average number of weeks of administration was 3.9.

Software Applications and Quality Checks
In-house Internet-based software applications on the Hospital-Based Cancer Registry and NCDIR Web sites were modeled for data capture, checking at data provided at entry and, subsequently, tracking patient follow-up, updating treatment information, and recording follow-up details. Collaborating centers were provided exclusive login IDs and passwords with instructions for data entry and transmission. The data were downloaded periodically at the NCDIR. Data checks included checking dates and verifying discrepancies in clinical information (Data Supplement). Lists of incorrect or unlikely cases were sent to the appropriate centers for rectification. In addition, a center-wise random sample of 10% of the cases was created, and centers were asked to re-abstract the medical records for certain essential parameters.

Statistical Analysis
Kaplan and Meier 20 and Cox proportional hazards ratio 21 analyses in the SPSS software package (SPSS, Chicago, IL) were used to calculate the 5-year cumulative survival (FCS) percentages and fatality risk (with statistical significance), respectively. Multivariable analysis was performed by using Cox proportional hazards regression analysis.   Table  2. The relative proportions in the RT parameters such as teletherapy dose, teletherapy plus brachytherapy combination, number of RT fractions, deciphered optimal RT (based on these three factors), type of RT machine, and RT fields suggest that those who received concurrent CT received better delivery of RT. In the first instance in Table 3, optimal RT was adjusted for RT machine and RT field; only RT field remained significant. Data in Table 3 show that patients who received cisplatin at 150 mg or more had better survival, although there was no survival difference in subgroups who received a total dose of more than 150 mg. Patient data for those who received optimal RT were adjusted compared with data for those who received a total dose of at least 150 mg cisplatin.

RESULTS
Patients were further regrouped according to combinations of optimal and/or suboptimal RT and total dose of cisplatin (Table 4 and Fig 1). The best survival was seen in patients who received optimal RT and 150 mg or more of cisplatin (FCS, 71.5%) and in those who received suboptimal RT with 150 mg or more of cisplatin (FCS, 76.2%). The shortest survival (FCS, 43.2%) was seen in the group of patients who received suboptimal RT with no CT. . Although an increased complication rate was observed with larger dose, there was no correlation between mortality and dose of cisplatin. The complications recorded were mainly parametrial fibrosis followed by hematologic, GI, and renal complications and skin reactions. These were largely comparable in the two treatment groups. Table 5 compares survival in our study with that in other key publications. The major differences between this study and the others are the sample size (substantially higher in this study) and the study setting (ie, this is the only study from a developing country). Otherwise, the survival benefit of RTCT over RT is comparable. Figure 2 depicts the Kaplan-Meier comparative survival curves for patients who received RT alone and those who received RTCT.

DISCUSSION
The survival benefit of RTCT over RT alone was reported in clinical trials in 1999. [22][23][24][25][26][27] Since then, and with the alert issued by the National Cancer Institute, 28 several studies have validated the improved disease-free survival and overall survival of RTCT over RT especially in locally advanced cervical cancer. These studies include updates, systematic reviews, and meta-analyses. 19,23,29 A recent Chinese study 30 reported a 5-year overall response rate of 67% in the RTCT arm and 53% in the RT arm. Singh et al 31 reported less morbidity and mortality with neoadjuvant CT. There have been some contradictory reports. 14,32 But the authors 14 concluded that bulky tumors, poor nutritional status, and small sample size could have contributed to differences in outcome not being observed.
The current focus of clinical research has shifted to determining the efficacy of other drugs and examining aspects of tolerance, toxicity, and effectiveness of a lower dose of cisplatin. Ushijima et al 33 reported a favorable response with an average total dose of 200 mg of cisplatin. We also did not observe significant differences in survival with total cisplatin doses of more than 150 mg. The cisplatin dose compared with RT dose 17 requires further investigation, at least in the Indian context wherein nutritional status, immunity, and comorbid conditions could play a role.
When the analysis was performed by using optimal and/or suboptimal RT and 150 mg total dose of cisplatin, we found that even patients who had received suboptimal RT but with a total dose of cisplatin of 150 mg or more did as well as or better than those who received optimal RT, suggesting that cisplatin is more important than optimized RT.
Improved survival with RTCT over RT in early-stage (IA2-IIA) cervical cancer has been reported. 34,35 Surgery and not RTCT had an impact on survival in this study of early-stage cancer.
Most reports of superior survival with RTCT are from developed regions, although a majority of Several factors could contribute to patients not receiving CT, such as the lack of qualified/trained clinical oncologists and lack of awareness about the need for such therapy. The data on comorbid conditions and performance status do not suggest that renal insufficiency is a factor in not administering cisplatin. A majority of the patients in this study underwent treatment in established, wellequipped cancer hospitals in which administering chemotherapeutic agents on an inpatient or outpatient basis is a regular daily affair. The reasons for not administering cisplatin probably do not relate to the medical condition of the patient or the facilities available at a given center. Rather, those reasons probably relate to patients' problems such as cost of treatment, the difficulties involved in repeated hospital visits (including long-distance travel), and a false sense of doing well after initial treatment. Data from HBCRs 37 show that until the year 2000, less than 10% of patients with locally advanced cervical cancers received RTCT. As of 2011, the percentage was 57.9%, but in a medical college HBCR it was 10%.

Limitations
Pooled data from different types of institutions (cancer centers, medical colleges, private hospitals) has the advantage of representing the complete spectrum of patients and treatment but the disadvantage of having subjective information on some variables that cannot be adjusted in analysis. Standards of care can vary from center to center. There is no selection bias because all patients who  received cancer-directed therapy in their respective institutions have been accounted for, and exclusion criteria are based on scientific logic. However, it is possible that a few patients received additional treatment elsewhere which, because of challenges in clinical follow-up, could not be quantified. There was no regular clinical follow-up per National Comprehensive Cancer Network guidelines. 38 Therefore, details of toxicity, recurrence, and disease-free survival and/or progression-free survival could not be accurately ascertained.

Strengths and Opportunities
This study is a foremost example of cancer registration because of its national program that evaluates and provides critical findings that could have an impact on patient care. Dynamic Internet-based data capture, data checks, and analysis had several advantages in ensuring standard quality data. NCDIR, a nonprofit organization with software experts as full-time faculty is unique, and along with its medical and statistical expertise, it has the strength of clinical neutrality, a distinct feature spelled out in its manifesto. The NCDIR research panel on cancer and scientific advisory committee have reputed oncologists from all subdisciplines. This POCCS is now an intramural activity of NCDIR. Thus, all 27 regional cancer centers and many other institutions have joined this study, providing an exceptional opportunity to examine, evaluate, and redesign treatment management in cervical and other cancers. A network of cancer hospitals linked to a central coordinating center with a system for accruing good clinical data through modern electronic information technology is in place. This number-38,771-is conservative because the results presented here are based on treatment provided in some renowned cancer treatment centers that are well equipped in terms of both facilities and skilled staff. In one participating medical college, the proportion of patients who received RTCT was less than 10%. If that proportion is used, then 90% of 75,873, or 68,286 cancers of the cervix, would not receive RTCT.
Perhaps this scenario in standards of care may be no different in other less developed regions of the world, where the majority 39 of cervical cancers (84.3%) occur. The global burden for 2012 was 527,600. 1 There seems to be a need to reiterate the NCI 28 alert that advised supplementing chemotherapy to radiation for locally advanced cancer of the cervix in the context of the developing world. If governments included cisplatin in the list of essential drugs, that would greatly increase its availability and at an affordable cost.