Clinico-pathological Profile of Lung Cancer at AIIMS : A Changing Paradigm in India

Lung cancer is one of the commonest cancers and cause of cancer related mortality worldwide. It accounts for 12.7% of all new cancer cases and 18.2% of all cancer related deaths, throughout the world (Parkin et al., 2008). In India, it is the commonest and most lethal cancer among males accounting for 10.9% of all cancer cases and 13% of cancer related mortality (Parkin et al., 2008). In general, the incidence of lung cancer throughout the world reflects the prevalence of smoking, and patterns of lung cancer appear attributable to the type of smoking (Valaitis et al., 1981). In recent years, there has been a great interest in the histological characterization of lung cancer in view of newer histology guided therapeutic modalities and genomic classification of lung carcinoma (Standfield et al., 2011; Scagliotti et al., 2011). Broad classification of lung cancer into NSCLC and SCLC, is being challenged (Ettinger et al., 2012). In western countries and most of the


Introduction
Lung cancer is one of the commonest cancers and cause of cancer related mortality worldwide. It accounts for 12.7% of all new cancer cases and 18.2% of all cancer related deaths, throughout the world (Parkin et al., 2008). In India, it is the commonest and most lethal cancer among males accounting for 10.9% of all cancer cases and 13% of cancer related mortality (Parkin et al., 2008). In general, the incidence of lung cancer throughout the world reflects the prevalence of smoking, and patterns of lung cancer appear attributable to the type of smoking (Valaitis et al., 1981).
In recent years, there has been a great interest in the histological characterization of lung cancer in view of newer histology guided therapeutic modalities and genomic classification of lung carcinoma (Standfield et al., 2011;Scagliotti et al., 2011). Broad classification of lung cancer into NSCLC and SCLC, is being challenged (Ettinger et al., 2012). In western countries and most of the 1 Asian countries, adenocarcinoma has surpassed squamous cell carcinoma (Valaitis et al., 1981;Janssen-Heijnen et al., 2003). This shift seems to be attributable partly to the changed smoking pattern and increasing incidence of lung cancer in females and non smokers (Valaitis et al., 1981;Thun et al., 2006;Wakelee et al., 2007).
Most of the older and some recent Indian series have described squamous cell carcinoma as the commonest histology (Behera et al., 2004;Singh et al., 2010). Population based cancer registries in India give information of site wise distribution of cancer and histological subtypes are not taken into consideration. For such information we need large hospital based data. There is an immense need of information about the pattern of histological subtypes and clinical characteristics of the disease in the recent time because of paucity of such data in India. We have analyzed clinical and pathological profile and treatment outcome of lung cancer patients treated at our centre in last 3 years, which is a regional cancer centre registering approximately 8,000 new cancer cases every year.

Materials and Methods
We analysed 434 consecutive, pathologically confirmed lung cancer patients, registered at our centre in Lung Cancer Chemotherapy Clinic between July 2008 and June 2011 (a period of 3 years). Clinical, demographic, treatment and outcome related informations were collected from the case record files and entered in a predesigned proforma. The available histopathology slides were retrieved from the archives and were reviewed by an independent pathologist. Slides were reviewed on the basis of morphology and available IHC (done at the time of initial reporting). Restaging was done according to AJCC staging system 7 th edition based on the available clinical and radiological findings. For treatment outcome and survival analysis only those patients were included who had received at least one treatment modality i.e. surgery, radiotherapy (radical or palliative), chemotherapy (at least one course of chemotherapy) or tyrosine kinase inhibitors (TKIs). We excluded patients with incomplete clinical informations.
Patients were treated as per the departmental treatment policy with multidisciplinary team approach. Responses were ascertained clinically and radiologicaly and coded according to revised RECIST criteria (Eisenhauer et al., 2009). Patients, after completion of treatment or on oral TKI, were followed every 3 monthly or early if clinically indicated. Detailed physical examination and chest X ray were done on each visit. CECT or other imaging modality (PET-CT) was repeated every 3-6 monthly or early if indicated. Patients with disease progression were offered second line treatment or supportive care only as per patient's performance status and preference.
Patients were considered on continuous follow up if the last visit was within 3 months of data censoring. In cases where last visit was more than 3 months ago, attempts were made to contact the patient by telephone and/or a reply letter. Patients were followed from the date of registration to the date of death and were censored at the date they were last known to be alive i.e. date of last follow up (if lost to follow-up) or May 31, 2012, whichever came first.
Permission for viewing case records was obtained from the institute's committee and the medical record department. Confidentiality of the patient's identity was maintained.

Statistical analysis
The data was censored at 31 st May 2012 or last follow up date (if lost to follow up). Descriptive statistics was used for describing demographic and clinical characteristics. Survival was estimated by the Kaplan-Meier method and the Cox Regression Model (univariate and multivariate) was used to identify significant prognostic factors. Factors which had p value <0.25 in univariate analysis were subjected to multivariate analysis. Analysis was done using the SPSS ver. 15(SPSS Inc, Illinois, Chicago).

Results
A total of 434 confirmed lung cancer cases were available for clinico-pathological analysis during the study period. The median age of the study population was 55 years (23-84years). Majority of the patients were between 50-70 years of age. Male to female ratio was 4.6:1. Out of the total population, 295 patients (67.97%) were smokers (active or former). Bidi (an indigenous form of tobacco) was the commonest mode of smoking (55.25%). There were 370 (85.25%) cases of NSCLC and 64 (14.75%) cases of SCLC. The diagnosis were predominantly based on biopsy (66.13%) followed by FNAC ( the commonest histology (37.3%). Two patients were found to have metastatic lung lesions (one squamous and one adenocarcinoma) and excluded from further analysis. Similarly one patient was found to have atypical carcinoid (initially reported as small cell carcinoma) and excluded from analysis. Table 2 summarizes the pathological profile before and after review.

Clinical presentation
Most common presenting symptom in both NSCLC and SCLC was cough (57.03% and 57.8% respectively) followed by chest pain (51.89% and 48.44%). Fever was present in 20.54% patients of NSCLC and 15.63% of SCLC. Clinical features of SVC obstruction were present in 5.42% of NSCLC patients and 25% of SCLC patients respectively. Pleural effusion was present in 26.76%  like poor PS, significant co morbidities, logistics and patient's preferences. Baseline demographic details according to main pathological subtypes are summarized in Table 1.

Pathology review
We retrieved 174 slides for review from our archives. On review there was discordance from the initial report in 42 patients (24.13%). On review 24 patients of unclassified group could be reclassified into squamous cell carcinoma (10), adenocarcinoma (12) and large cell carcinoma (2). One patient with adenocarcinoma and one patient with squamous cell carcinoma were found to be poorly differentiated and regrouped with NSCLC NOS. Nine patients who were initially reported to have squamous cell carcinoma but found to be adenocarcinoma after review. Among the biopsy slides which were subjected to independent review, squamous cell carcinoma was the commonest histological subtype (33.33%) as per the initial report, but after review adenocarcinoma was found to be

Survival analysis and prognostic factors
All patients were treated with multidisciplinary approach according to patient's characteristics. Median follow up time for patients with NSCLC and SCLC were 14.9 months and 11.5 months. The median overall survival of patients with NSCLC was 12.8 months (95%CI 11.0-14.7) and the median progression free survival was 7.8 months (95%CI 6.1-8.8). On the other hand, the median overall survival of patients with SCLC was 9.1 months (95%CI 6.8-11.4) and the median progression free survival was 6.8 months ). The Kaplan Meir survival curves are shown in Figure 1.
We analyzed patient's characteristics for their impact on survival in NSCLC and SCLC separately by univariate and multivariate cox regression analysis. Multivariate analysis identified PS>2, poorly differentiated/unclassified histology and serum albumin <3.5 mg/dl as significant adverse risk factors for overall survival while for progression free survival only PS>2 was significant among patients with NSCLC (Table 3). For SCLC none of these factors was found to be significant on multivariate analysis (Table 4).

Discussion
The aim of this analysis was to study the current clinico-pathological profile of lung cancer patients at our centre and to assess their outcome. A total of 434 confirmed lung cancer patients who were diagnosed and treated over a period of 3 years were analyzed. As compared to Western population, median age of our patients was a decade younger (Blanchon et al., 2006;Cetin et al., 2011;Albain et al., 1991). Most of the previous studies Indian have reported the similar median age (Behera et al., 2004;Singh et al.,2010;Prasad et al 2004;Khan et al.,2006). Smoking is associated with most of the lung cancer cases (Vineis et al., 2004;Giovino et al.,2002). In our study, we found that up to 70% patients were smokers (current or former). Majority were bidi smokers which is an indigenously prepared, unfiltered and crude form of tobacco smoking prevalent mainly in the rural population. Previous Indian series have shown that majority of patients with lung cancer were bidi smokers (Behera et al.,2004;Prasad et al 2004).
In our study, we found adenocarcinoma to be the commonest histological subtype, accounting for 39% of all lung cancer cases. Over last few years there has been a shift of histological profile towards adenocarcinoma worldwide (Valaitis et al., 1981 ). However, most of the Indian series still report squamous cell carcinoma to be the commonest subtype (Behera et al., 2004;Singh et al.,2010;Prasad et al 2004;Khan et al.,2006). As a result of recent evolution in newer histology based treatment approaches, there is a need of proper histological sub typing of lung cancer (Scagliotti et al., 2011;Cooper et al.,2011). In our study we observed that up to 21% of lung cancer cases were still labelled with the generic term of NSCLC. However, on independent review, this proportion was reduced to 8% and we could reclassify many cases from the unclassified group. This signifies the critical role of pathology review in lung cancer in the present era of personalised treatment. Field et al (2004) have previously shown the relevance of independent pathology review from SEER database. Interestingly, most common histological subtype among the slides which were reviewed was squamous cell carcinoma according to the initial report, but on review adenocarcinoma turned out to be the commonest subtype. This under diagnosis may be one reason of reporting Asian Pacific Journal of Cancer Prevention, Vol 14, 2013 493 DOI:http://dx.doi.org/10.7314/APJCP.2013.14.1.489 Clinico-pathological Profile of Lung Cancer at AIIMS: A Changing Paradigm in India lower frequency of adenocarcinoma by most of the Indian studies. An independent central review can reveal the actual pathological profile of lung cancer in countries like India where squamous cell carcinoma is still reported to be higher in frequency in contrast to the developed world.
Most of the patients in our study had advanced disease (stage IIIB-IV) at the time of presentation (70% in NSCLC and 71% in SCLC). In the series from west as well as from India, it is reported that 50-70% cases of NSCLC and up to 2/3rd of SCLC usually present in advanced stage (Behera et al., 2004;Blanchon et al., 2006;Govindan et al 2006;Grivaux et al 2011).
It is a well known fact that not all patients with lung cancer are able to receive anti cancer treatment. In our study 28.57% patients (29.46% of NSCLC and 23.44% of SCLC) didn't receive any anticancer treatment. These figures varies from 19% in USA, 33% in Australia, 37% in Scotland, and 50% in Ireland and New Zealand (Fry et al., 1999;Vinod et al., 2008;Erridge et al., 2008;Mahmud et al., 2003;Stevens et al., 2007). Vinod et al (2010) have addressed this issue in their study and found that in actual practice 20% patients didn't receive any treatment; however guideline recommendations for no treatment were there in only 4% patients. On multivariate analysis, it was observed that main factors responsible for this discrepancy were older age, poor PS, NSCLC histology and social reasons.
Among NSCLC, only 5.36% patients underwent surgery. This proportion appears to be quite low considering the fact that we had about 25% cases of NSCLC with stage I-IIIA. If we look at the proportion of patient undergoing surgery out of early stage (operable) disease, then it would be 15% (14/93). Even studies from United States demonstrate that 30-60% of cases of early stage lung cancer don't undergo surgery. They found racial differences in proportion of patients undergoing surgery, apart from disease related factors and co morbidities (Esnaola et al., 2008;Cykert et al., 2010). In our study, 39.08% of NSCLC and 36.7% of SCLC patients received radiotherapy to the primary site, out of which only 29.4% (of patients receiving radiotherapy) of NSCLC and 27.7% of SCLC, received radiotherapy in radical doses. There was marked underutilization of this treatment modality. Generally guidelines recommend that up to 76% of lung cancer patients should receive radiotherapy at some time point during their treatment. However in actual practice, 36-70% cases actually receive radiotherapy, worldwide (Delaney et al., 2005).
Median overall survival and median progression free survival of patients with NSCLC were 12.8 months and 7.5 months respectively. This reflects the large proportion of advanced stage disease in the study population. Furthermore, many of early stage patients could not complete the intended treatment and hence had inferior survival than expected. On multivariate analysis we found that PS>2, poorly differentiated histology and serum albumin levels <3.5mg/dl were associated with inferior survival however for progression free survival only PS>2 was found to be significant adverse risk factor. In a larger study of 4669 patients, Blanchon et al (2006) have developed a prognostic score for NSCLC and demonstrated that age, stage, PS, histology and treatment modalities used had a significant impact on survival. In SWOG (South-West Oncology Group) experience of advanced NSCLC, PS, female sex, age, haemoglobin and serum LDH were found to be independent prognostic factors (Albain et al., 1991).
The outcome of SCLC in our study was poorer than that of NSCLC. The median survival of SCLC patients was 9.1 months and median progression free survival was 6.8 months. These figures are consistent with the literature (Albain et al., 1990). On multivariate analysis none of the factors was found to be significant. This may be due to small number of patients among SCLC group. Mohan et al (2006) have previously demonstrated in their study that, PS, stage, Hb and symptom burden were independent prognostic factors for survival. Analysis of SWOG database also shows that age, PS, LDH, female sex were independent factors for survival for limited disease while only LDH was significant for extensive stage disease (Albain et al., 1990).
In conclusion, there is a paradigm shift in the clinico-pathological profile of lung cancer in India. Adenocarcinoma may be the commonest histological subtype in this part of the world also, provided a careful independent pathology review is done. Analysis of a larger cohort from multiple institutions would reflect the true pattern. Outcome of these patients still remain poor because of presentation in advanced stage and poor PS and many of them are not able to receive adequate treatment.