Indian Journal of Cancer Medknow Publications on behalf of Indian Cancer Society ISSN: 0019-509X EISSN: 1998-4774 Vol. 48, Num. 2, 2011, pp. 181-186

Indian Journal of Cancer, Vol. 48, No. 2, April-June, 2011, pp. 181-186

Original Article

Utility of PET in unknown primary with cervical metastasis: A retrospective study

MR Dandekar¹, S Kannan¹, V Rangarajan², NC Purandare², DA Chaukar¹, A Deshmukh¹, AK D«SQ»cruz¹

¹ Department of Head and Neck Surgery, Tata Memorial Hospital, Parel, Mumbai, India
² Department of Bioimaging, Tata Memorial Hospital, Parel, Mumbai, India

Correspondence Address: A K D'cruz Department of Head and Neck Surgery, Tata Memorial Hospital, Parel, Mumbai India docdcruz@gmail.com

Code Number: cn11046

DOI: 10.4103/0019-509X.82882

Abstract

Keywords: Cervical neoplasm, PET scan, unknown primary neoplasm

Introduction

Carcinoma of unknown primary with cervical metastasis is rare accounting for 2-5% of all head and neck cancers. ^[1] Identification of primary tumor has an influence on the treatment and potentially on patient prognosis ^[2] thus justifying the extensive workup for detecting primary in suspected unknown primaries. The development and application of more sophisticated diagnostic tests will increase the yield of the primary tumors. ^[3] Today, FDG-PET is recommended as an investigation in unknown primary tumors according to the NCCN 2010 guidelines. ^[4] Many studies are published on the use of FDG-PET in detection of primary in patients with cervical metastasis. ^{[5],[6],[7],[8],[9],[10],[11],[12],[13],[14]} Unfortunately, most of the studies have a very small sample size. Moreover, the routine use of FDG-PET is limited since it is an expensive investigation with limited availability. There is no Indian data studying the diagnostic work up of patients with unknown primary tumors with cervical metastasis. This retrospective study attempts to determine the utility of FDG-PET over conventional imaging modalities in such cases.

Materials and Methods

Case records of all patients of cervical metastasis without any clinically detected primary tumor were retrieved from the nuclear medicine database and the medical records department from June 2005 to July 2009 of a tertiary level oncology institute. Provisional diagnosis of unknown primary was made when a thorough clinical examination including endoscopic evaluation of potential primary sites revealed no primary tumor. Clinical examination encompassed a complete head and neck examination including the examination of the neck node, scalp, ear, the oral cavity, a 70° endoscopic assessment of the oropharynx, larynx, and hypopharynx as well as a 0° endoscopic assessment of the nasal cavity and the nasopharynx. Examination was performed by an experienced head and neck surgeon. Every patient was discussed in a multidisciplinary clinic. Pathological confirmation of the neck node was made by a fine needle aspiration cytology (FNAC). The imaging modality deemed appropriate by the treating clinician was advised. Most of the patients already had some imaging, either a computed tomography (CT) scan or magnetic resonance imaging (MRI) with them during their first visit, this being a referral center. In addition, there were some logistic implications in view of the volume of patients and the limited availability of FDG-PET. This led to an inherent heterogeneity among the patients with respect to the imaging modality used.

Since most patients had a CT scan done before they were referred to our center, the imaging protocol was not standard. Although most patients underwent CT scan of the head and neck area with contrast, there were nine patients with an additional CT scan of the thorax and one patient with CT abdomen and pelvis, all being contrast enhanced. When the patient underwent CT scan in our center, the protocol followed was that of a CT scan of the head and neck up to the mediastinum with intravenous contrast. Only one patient underwent a MRI from a center outside our institute. It was performed with multiplanar spin echo T1- and T2-weighted images and post-contrast T1-weighted sequences of the neck. All patients underwent a FDG-PET scan or a FDG-PET CT scan at our institute.

Rationale for using FDG PET

Malignant lesions show increased expression of glucose transporters (GLUT-1) and hence FDG is known to concentrate in malignant lesions.

FDG-PET is used to unmask the possible site of primary which is not seen clinically or on morphological imaging.

Patient preparation and PET-CT imaging protocol

All patients were asked to fast for 4-6 h prior to the study and blood glucose levels were checked and confirmed to be less than 150 mg/dl. The studies were performed 1 h following intravenous administration of 370 MBq (10 mCi) of ¹⁸ F-FDG. Imaging was performed on a Discovery ST PET-CT system (GE medical systems).

It combines a 16 slice CT scanner with a dedicated PET (BGO plus crystal, dimensions 3.8 mm × 3.8 mm × 3.8 cm).

Image reconstruction and interpretation

Image fusion was performed using co-ordinate based fusion software and subsequently reviewed at a workstation (Xeleris) that provided multiplanar reformatted images and displayed PET images, CT images, and PET-CT fusion images.

Studies were interpreted independently by a nuclear medicine specialist and a radiologist. The CT data were used for anatomical localization and corroboration of the PET findings. Abnormal increase in FDG uptake at loco-regional sites in the head-neck as well as the distant sites was noted. The maximum standardized uptake values (SUVs) were automatically generated. Visual and semi-quantitative estimate of FDG uptake was used to classify findings as positive or negative.

The patients who underwent CT scan or MRI outside our institute had their images reviewed by an experienced radiologist. The criteria for positivity were the presence of an enhancing mass lesion, wall thickening, mucosal irregularity or ulceration, or any such morphological abnormality.

A total of 352 case records were retrieved. Inclusion criteria included all patients with a palpable neck node which was cytologically confirmed to be malignant (excluding lymphomas) without any clinically detectable primary tumor. Exclusion criteria were incomplete information in the case records, patients with previous treatments to the head and neck, patients who received treatment with palliative intent during the first visit in view of poor performance status or large fungating nodes and patients with parotid lymphadenopathy. A total of 240 patients were excluded [Table - 1]. Thus, 112 patients were included in the study. Data extracted from the case records included age, sex, neck node level, neck node staging, conventional imaging findings, and findings on FDG-PET. Neck node staging was according to the 2010 American Joint Committee on Cancer staging criteria (7th edition). ^[15] Patients were divided into three groups based on the type of imaging they underwent, i.e. either a conventional imaging modality, PET/PET CT scan or both. Conventional imaging modality included a CT scan or a MRI. All patients who underwent conventional imaging were included in group I (53 patients) and group II included patients who underwent FDG-PET (59 patients). Out of the 59 patients in group II, a subgroup of 40 patients underwent conventional imaging in addition to FDG-PET which were included in group III.

Statistical analysis

Data were entered and analysed using SPSS version 14.0. True positive was considered when there was either histopathological confirmation or if there were corroborative findings which were considered to be significant by the treating clinician or the nuclear medicine specialist who reported the FDG-PET findings.

False positive was considered when the imaging modality showed a primary tumor, but it did not corroborate with the clinical examination (including endoscopic evaluation) or histological diagnosis.

False negative was when the imaging modality did not show a primary tumor, but the histological diagnosis or tumor markers proved otherwise.

True negative was when the imaging modality did not show the suspected primary and it corroborated with the clinical findings, histological findings, and/or tumor markers. Sensitivity, specificity, positive and negative predictive values were calculated between groups I and II. Among the patients in group III, the added advantage of FDG-PET over conventional imaging was determined with its influence on change in the management plan. Association between distant metastasis and neck nodes was analysed using multivariate logistic regression analysis.

Results

Patient and tumor characteristics are as shown in [Table - 2]. A total of 91.07% of patients presented with nodes >3 cm (N2/N3 disease). Of the 53 patients in group I, 52 underwent CT scan and 1 patient underwent a MRI. Conventional imaging hinted toward a primary in 29 patients (54.7%) [Table - 3].

In group II (59 patients), 46 underwent FDG-PET CT and 13 underwent FDG-PET scan. Primary sites were suggested in 23 cases (38.98%), 11 out of which were infraclavicular primaries [Table - 4]. Among the 27 patients with low neck nodes, FDG-PET detected 12 primaries, 9 of which were infraclavicular (75%).

Conventional imaging showed a sensitivity and specificity of 92.3% and 50% whereas FDG-PET showed a sensitivity and specificity of 92.8% and 71.4%, respectively. The positive predictive value of FDG-PET was 56.5% and the negative predictive value was 96.15%.

FDG-PET detected metastasis in 31 patients (52.54%). This included infraclavicular lymph nodal metastasis in 12 patients, distant organ metastasis in 6 patients, and both in 13 patients. Eighteen patients with distant metastasis had low level lymph nodes (58.06%). One patient presented after a modified neck dissection; hence, the nodal level could not be determined. Twenty-six patients with distant metastasis presented with N2, N3 nodal disease (83.87%) [Table - 5]. Of the patients with multiple neck nodes at presentation (N2b, N2c), 16 patients had distant metastasis [Table - 6]. Multivariate logistic regression analysis was performed for distant metastasis and nodes including N2, N3 nodal disease, level of nodes, and multiplicity (N2b, N2c). It showed that distant metastases were associated significantly with multiple nodes at presentation (P = 0.007).

The authors attempted to determine the predictors of distant metastasis. The factors considered were detection of primary and (standardised uptake value) SUV _max . However, no association was established.

Group III, in which 40 patients underwent both FDG-PET and conventional imaging, FDG-PET added information to conventional imaging in 32.5% patients with respect to primary and metastasis (8 by detecting metastasis, 4 by detecting primary, and 1 by detecting both primary and metastasis).

FDG-PET brought about an overall change in management in 38.9% of patients [Table - 7].

Discussion

Conventional imaging modalities such as CT/MRI detect anatomical abnormalities. FDG-PET detects metabolically active sites which may not have any significant anatomical abnormality [Figure - 1]. This has significance in unknown primary tumors in which one of the postulated theories is that there is a small primary which is missed clinicoradiologically. ^[16] There are many studies comparing FDG-PET with conventional imaging. It was found that FDG-PET has a significantly higher sensitivity ranging from 66 to 87% as opposed to CT (43%). ^[5],[17] Dong et al,^[18] in their meta analysis reported a pooled sensitivity of 81% and a specificity as 82% whereas Kwee ^[19] et al. in a recent meta analysis of 11 studies reported a pooled sensitivity and specificity as 84%. Our study showed that both conventional imaging and FDG-PET have similar sensitivity; however, FDG-PET is more specific compared to CT (71.4% as opposed to 50%).

There are two studies from India on unknown primary with cervical metastasis. ^[20],[21] Both these studies analysed the outcome of unknown primary tumors. However, no comparison was made on the imaging modalities or the diagnostic work up. In our study, the inherent heterogeneity in the imaging performed on the patients helped us to compare the imaging modalities in similar clinical presentations. The majority of the patients had some conventional imaging done prior to their first visit as this was a referral center. It was thus possible to compare additional utility of FDG-PET over conventional imaging due to the subgroup which had both the imaging modalities.

The majority of patients with unknown primaries have nodes larger than 3 cm, i.e. N2, N3 disease at presentation as opposed to those with a known primary because the patient is asymptomatic before appearance of a neck swelling. ^{[5],[22],[23]} In our study, 91.07% of patients had nodes larger than 3 cm. A study by Liao et al.^[24] suggested that clinical N2, N3 disease, pN2, the presence of more or equal to five positive lymph nodes, Level IV, V, and N2c disease were statistically significant risk factors for the 5-year distant metastasis rate in oral cavity squamous cell carcinoma. It was an important observation in our study that of the 31 patients with distant metastasis diagnosed on FDG-PET only five patients had an N1 disease. However, on comparing multiplicity of nodes, bulky nodal disease and the level at presentation, this study showed a significant association only between distant metastasis and number of nodes at presentation (N2b, N2c) (P = 0.007). Lim et al.^[25] determined in their retrospective analysis of oral and oropharyngeal squamous cell cancers in a multivariate analysis that the pathologic positive lymph node especially bilateral neck metastasis was an independent risk factor for the appearance of metastasis whereas a study from Memorial Sloan-Kettering Cancer Centre by Matsuo et al. showed that clinical nodal stage was an independent risk factor for distant metastasis in laryngeal cancers. ^[26]

It is known that higher FDG-PET SUV is associated with poorer outcome in head and neck cancer. ^[27] Hence the authors attempted to determine predictability of distant metastasis in patients with high SUV max; however, no association was established.

The identification of distant metastasis mainly helped in changing the management intent from curative to palliative in up to 18 patients. An overall change in management was brought about in 38.9% of patients. Literature reports an overall therapeutic change ranging from 25% ^[22] to up to 53%. ^[28]

Lower cervical lymph nodes (Level IV and supraclavicular) are usually associated with an infraclavicular primary. In our study, among the group II patients, 27 patients had low neck nodes. FDG-PET detected primary in 12 of these patients, 9 of which were infraclavicular.

This study is not without drawbacks. It is a retrospective study and hence had all the shortcomings associated with one. If FDG-PET detected a primary lesion in the head and neck which would be covered by the same radiotherapy fields as the neck node, no attempt was made at confirming the primary tumor. This excluded nasophayngeal tumors which required a separate field from that of the nodes. In addition, if the diagnosis of metastasis on FDG-PET corroborated with the clinical findings such as the nodal stage and level or if the metastasis was from anatomically difficult areas such as the mediastinum or bone, it was not histologically confirmed. This could be critically looked upon as the absence of a gold standard. However, the decision of the clinician could be clinically justified. This study could have led to an overestimation of the diagnostic performance of PET, because it defined unknown primary tumors following clinical examination and routine biochemical investigations and a X-ray of the chest as opposed to biochemical and clinicoradiological definition of unknown primaries according to Pavlidis et al. ^[29] However, the aim of the study was to assess the advantage of a relatively new imaging modality (PET) to the conventional imaging. Hence, the authors were justified in considering only the clinically diagnosed unknown primaries. This study did not correlate detection of primary tumors with patient outcome due to limited followup. Another drawback of this study is the limited number of patients in each group. However, the number is significant compared to other studies on unknown primary tumors.

Conclusion

The majority of the patients with cervical metastasis with unknown primaries present as N2/N3 disease. PET has an additional value over CT because it aids in screening infraclavicular primaries and detects distant metastasis thus influencing management intent. However, the benefits of FDG-PET are dampened by the affordability and availability of this investigation. A cost-effective analysis would help in determining whether FDG-PET is cost-effective over a battery of tests which are performed for diagnosis of primary lesion.

Acknowledgments

The authors thank Dr. Gouri Pantvaidya and Dr. Priya Ranganathan for their contribution to the statistical analysis.

References

Copyright 2011 - Indian Journal of Cancer

The following images related to this document are available:

Photo images