The standard treatment for patients who have carcinoma of an unknown primary has not been established to date. The prognosis for patients with carcinoma of unknown primary is poor. As a group, the median survival is approximately 3-4 months with less than 25% and 10% of patients alive at 1 and 5 years, respectively. Materials and Methods: In this retrospective analysis of 104 patients with carcinomas of unknown primary site, those treated with radiotherapy, chemotherapy or a combination of both was included. Radiation dose varied from 8 Gy in a single fraction to 60 Gy in 30 fractions over 6 weeks. In chemoradiation, cisplatin 30 mg/m 2 was given concurrently with radiation dose of 40 Gy in 20 fractions over 4 weeks. Chemotherapy regimens used were cisplatin-based in nine patients and adriamycin-based in five patients. All drugs were administered intravenously on a once in every 21-day cycle. Results: Twenty-eight patients (28%) presented with two or more sites involved. The pathologic diagnosis was adenocarcinoma (30 patients; 30%) squamous carcinoma (30 patients; 30%) and metastatic carcinoma not otherwise specified in 26 (26%) patients. In patients with chemotherapy, four patients had complete response and four had partial response. Disease stabilization was seen in three patients. In patients treated with chemoradiation, complete response was seen in ten patients, these patients had neck node metastases with squamous histology mainly (seven patients). The median response duration was 7 months and the median time to progression was 9 months. Median overall survival was 6 months. Grade 3 and 4 haematological toxicities were more frequent in the patients treated with cisplatin and adriamycin chemotherapy. Conclusions: In patients with an unknown primary, squamous histology responds better as compared with adenocarcinomas. Patients with squamous histology can be cured with chemoradiation even with neck node metastases. However, the majority of patients are suitable only for palliative treatment because of advanced disease and poor general condition.
Keywords: Carcinoma, neck nodes, unknown primary
The standard treatment for patients who have carcinoma of unknown primary site has not been established to date. Carcinomas of unknown primary site represent a distinct but heterogeneous group of metastatic tumours accounting for 2-4.5% of all cancers.  At our institute, they comprise about 2.9% of all malignancies per year. These patients pose a problem because of their poor general condition at presentation secondary to disseminated disease and poor tolerance for therapy. Hence, in a majority of such patients, the aim is to provide them good palliation with minimum morbidity. Only a chunk of such patients are fit for radical treatment.
Between January and December 2003, a total of 104 patients with unknown primary were registered in the Department of Radiotherapy at our institute. These patients had histologically or cytologically documented metastatic carcinoma. Complete history was taken and thorough physical examination was done. Blood chemistry including serum tumour markers for prostate-specific antigen (PSA), α-fetoprotein (α-FP) and β-human chorionic gonadotropin (β-HCG) in men, urine analysis, mammography and CA-125 in women, thoracic and abdominopelvic computerized tomography scans, bone scan, symptom and sign oriented imaging and panendoscopy with multiple biopsies were done to look for the primary. Light microscopic pathologic analysis showed a well or poorly differentiated adenocarcinomas, a poorly differentiated carcinoma or squamous cell carcinomas or metastatic carcinoma not otherwise specified.
Pathologic analyses were carried out to rule out lymphomas (staining for leucocyte common antigen), extragonadal germ cell neoplasms (α-FP and β-HCG), sarcomas (cytokeratin and vimentin), malignant melanoma (S100 and HMB45), neuroendocrine tumours (chromogranin and synaptohysin) and prostatic adenocarcinomas (PSA) in men.
Patients were treated with radiotherapy alone, chemotherapy or combination of both. Radiation dose varied from 8 Gy in single fraction to eight patients, 20 Gy/5 fractions/1 week to 11 patients, 30 Gy/10 fractions/2 weeks to 16 patients, 40 Gy/20 fractions/4 weeks to six patients, 50 Gy/25 fractions/5 weeks to five patients and 60 Gy/30 fractions/6 weeks to 17 patients. Palliative radiation was given to patients with large solitary lung lesions (two patients), multiple brain metastases (13 patients) and patients with painful bony metastases (eight patients) under steroids cover whenever required. Sixteen patients were treated with chemoradiation; cisplatin 30 mg/m 2 was given concurrently with radiation dose of 40 Gy/20 fractions/4 weeks. Chemotherapy regimens used were cisplatin-based in nine patients and adriamycin-based in five patients. All drugs were administered intravenously once in every 21-day cycle.
Assessment of response and toxicity
Objective tumour response was assessed after completion of the treatment according to WHO criteria: complete response (CR) was defined as complete disappearance of all objective evidence of disease for a minimum of 4 weeks. Partial response (PR) was defined as ≥50% decrease in sum of the products of diameters of measurable disease for a minimum of 4 weeks. Stable disease (SD) was defined as less than 50% decrease in the sum of the products of measurable disease or <25% increase with the evidence of progression for a minimum of 3 months. Patients were said to have progressive disease (PD) when new lesions appeared during study or an increase in existing lesions. Acute toxicity scoring was done during the treatment according to WHO  criteria for haematological, renal and gastrointestinal systems.
The primary endpoint was overall response rate (ORR). It was defined as the proportion or patients who achieved a CR or a PR confirmed by a second evaluation at least 1 month but generally 3 months later. Time to progression (TTP) was defined as the interval between the start of therapy and earliest date of disease progression. An increase of 25% or more in measurable lesions (single lesion or sum of the products of all measurable lesions), an estimated increase of the same magnitude in nonmeasurable disease or appearance of new lesions constituted evidence of progression. Discontinuation of treatment with documented evidence of clinical deterioration due to cancer or death due to cancer or death of unknown cause while receiving treatment or with in 6 weeks of therapy, also constituted disease progression.
Secondary endpoints included duration of overall response, time to response (TTR), number of deaths and overall survival. TTR was defined between start of therapy and earliest documentation of response. Duration of OS was defined as the interval between start of therapy and death for any reason.
Median TTP was estimated by Kaplan-Meier product limit method. TTP was censored in the following circumstances: patient was still receiving treatment without evidence of progression, patient died of unknown cause without evidence of clinical deterioration due to cancer and patient discontinued treatment for any reason without evidence of clinical deterioration due to cancer before discontinuation. ORR was analysed by logistic regression. The dominant site as assessed at baseline was defined as neck node if only neck node disease was present, as bone if skeletal metastasis was present (regardless of coexistent neck node disease) without involvement of visceral site and as viscera if visceral metastases were present (irrespective of neck node or bone involvement).
Patient characteristics were as shown in [Table - 1]. Seventy-two percent of the patients were males and 28% were females. Median patient age was 62 years.
Twenty-eight patients (28%) presented with two or more sites involved. Neck node was the commonest site of presentation, 36 (35%) followed by liver and ascites 22 (21%) each, brain 13 (12%), bone 8 (8%) and lungs 3 (3%). The pathologic diagnosis was adenocarcinomas (36 patients; 35%), squamous carcinoma (30 patients; 29%), metastatic carcinoma not otherwise specified in 20 (19%) patients and poorly differentiated in 18 (17%).
Among 104 patients, only 56 (54%) were treatable, rest were given palliative treatment in the form of single fraction or supportive treatment in the form of analgesics and steroids. In patients treated with chemoradiation, CR was seen in 14 (39%) patients; these patients had neck node metastases with squamous histology mainly in ten (28%) patients. Among these, 11 (30.5%) patients were disease-free at the time of last follow-up. PR was seen in 12 (33%), SD was seen in 6 (17%) and 4 (11%) patients had PD as shown in [Table - 2]. In patients treated with chemotherapy, CR and PR were seen in four (28.5%) patients each. SD and PD were seen in three (21%) patients each. The median response duration was 7 months and the median TTP was 9 months. ORR was seen in 34 (33%) patients.
In patients treated with palliative radiation, symptomatic relief was seen in 17 (53%) patients. Mean symptom relief duration was 2.5 months. In patients who were suitable only for supportive care, adequate pain control was achieved in 12 (53%) patients.
The overall tolerance of these patients to combination chemotherapy was poor. This was due to poor KPS <60 in 60 (57%) patients at presentation. In patients treated with chemoradiation, four cycles were completed by 69% of patients. Grade 3 mucositis was seen in five (31%) patients and radiation treatment was interrupted for more than 1 week in four (28%) patients. Chemotherapy was stopped in 25% of patients due to deranged renal functions and in one patient due to haematological toxicity. Grade 3 and 4 haematological toxicities were more frequent in the patients (5, 36%) who were treated with a combination of cisplatin and adriamycin chemotherapy.
Overall median survival was 6 months. However, in patients with neck nodes, it was 22 months. Median survival in patients who were given only supportive care was 3 months.
Carcinoma of unknown primary is represented by a heterogeneous group of diseases all of which present with metastasis as the primary manifestation. The overall prognoses for patients with carcinoma of unknown primary is poor. As a group, the median survival is approximately 3-4 months with less than 25% and 10% of patients alive at 1 and 5 years, respectively. , Most large studies have shown that carcinoma of the lung and pancreas are the most common primary carcinomas that initially present as carcinoma of unknown primary. Other common malignancies such as colorectal, breast and prostate cancers infrequently present as carcinoma of unknown primary. Although the majority of diseases are relatively refractory to systemic treatments, certain clinical presentations of carcinoma of unknown primary carry a much better prognosis. In each instance, distinct clinical and pathologic details require consideration for appropriate, potentially curative, management. ,
In this study, we analysed 104 patients with unknown primary; an overall response was seen in 34 (33%) patients with complete response rate only in 18 (17%) patients. Patients who presented with neck node metastases and squamous histology responded better to treatment as compared to other sites and other histological diagnosis. Out of 18 patients with complete response, 10 (55.5%) had neck nodes with squamous histology in 7 (39%), respectively. Median survival in these patients was 22 months. A histologic diagnosis of metastatic carcinoma in cervical nodes requires a meticulous examination of the upper aerorespiratory tract. Histologically, these tumours are usually squamous cell carcinoma as seen in our analysis also, but occasionally may be adenocarcinoma, melanoma or anaplastic tumours.
It will be difficult to compare the outcome of these patients with those from western countries where most of the patients are in good general condition and the overall burden of disease is less due to the readily accessible medical help. In India, many patients present very late with advanced disease and poor general condition when they are suitable only for palliation. This is reflected in the tolerance profile of these patients to chemoradiation and chemotherapy alone.
Several variables of significant prognostic importance have been identified by multivariate analysis. Lymph node involvement and neuroendocrine histology were associated with longer survival; male sex, increasing number of involved organ sites, adenocarcinoma histology and hepatic involvements were unfavourable prognostic factors.  Adrenal involvement has also been noted to be a poor prognostic finding. The nodal status is considered the most important prognostic factor. In fact, the prognosis seems equivalent to that observed in patients with overt primary and similar nodal stage.  For patients treated with neck dissection, other prognostic factors include the number of lymph nodes, grading and extracapsular extension.  Metastatic adenocarcinoma is generally associated with a poor prognosis. Approximately 2%-5% of patients with primary squamous cell carcinoma of the head and neck region will present with cervical adenopathy as the primary disease manifestation; about 10% of this group will present with bilateral adenopathy. The 3-year survival rate ranges from 35% to 59% when patients with squamous or undifferentiated tumours are treated with radical radiation therapy, surgery or both. 
In patients with unknown primary squamous histology responds better as compared with adenocarcinomas. Patients with squamous histology can be cured with chemoradiation even with neck node metastases. However, the majority of patients are suitable only for palliative treatment in the form of radiation, chemotherapy or supportive care because of advanced disease and poor general condition at presentation.
[Table - 1], [Table - 2]