Lung cancer remains the most common cause of cancer mortality worldwide, despite advances in diagnosis, staging and therapeutic approaches.1,2

Non-small cell lung cancer (NSCLC) accounts for 75–80% of all cases.3 Since surgery offers the best opportunity for long-term survival and cure,4 appropriate staging is critical in determining treatment options. It is crucial to select suitable patients who have resectable disease and to avoid unnecessary surgery in advanced disease.4,5

NSCLC staging is based on the TNM system, using the tumor (T), node (N) and metastasis (M) evaluations, provided by non-invasive and invasive procedures.4,5

Traditionally, conventional non-invasive procedures include thorax and upper abdomen computed tomography (CT), with some patients requiring specific imaging.4,5 Previous data suggest, however, that staging was not accurate when based on these methods, with a former study demonstrating a slight concordance between clinical and pathological staging (21.7%, kappa=0.0418) in 60 patients with NSCLC submitted to surgery between 1999 and 2003.6

Despite advances in CT scan technology, the morphological information provided has limitations regarding tumor delineation from adjacent structures, limited sensitivity for microscopic disease and is frequently unable to discriminate lymph nodes that are enlarged owing to malignancy or to benign pathology.4,7

The possibility of quantifying the tumor uptake of fluorodeoxyglucose (FDG) and assess its metabolic activity has emerged, in the past two decades, with positron emission tomography (PET). Staging of NSCLC was one of the first approved indications for the use of PET.8 In the meantime, integrated PET/CT, providing both metabolic and anatomic information, has replaced stand-alone PET.9 Integrated PET/CT has the necessary anatomic detail to evaluate the tumor and distinguish malignant lesions from benign lesions with an accuracy of 82%, greater than either CT or PET alone.10 The same is true for lymph nodes evaluation (sensitivity (73%), specificity (80%), positive predictive value (PPV) (78%), negative predictive value (NPV) (91%), accuracy (87%))10 and for malignant extrathoracic metastasis detection (sensitivity (98%), specificity (92%), PPV (89%), NPV (98%)).11 Although PET/CT has demonstrated a good NPV value, it has a lower PPV, indicating that false positives (FP) may occur.10 Tissue sampling, using invasive non-surgical or surgical approaches, is therefore required, in order to confirm positive PET/CT results.

Besides its use for staging purposes, PET/CT also has a role in restaging patients after induction therapy, with literature data showing that it may be better than repeated CT or PET alone,12,13 and even superior to remediastinoscopy.13

PET/CT has become a fundamental technique for the preoperative evaluation of NSCLC and has led to changes in initial staging and treatment plans.14,15

The authors carried out a retrospective study with the main objective of comparing agreement between clinical and pathological staging in patients with NSCLC, operated throughout a period of 10.5 years, including a comparison between two subgroups, for the last of which PET/CT was available.

A retrospective analysis of clinical files of 150 patients with NSCLC, of the pulmonology department (lung cancer unit) of a university hospital, who underwent surgery over a period of 10.5 years (January 1999–July 2009), was carried out. The authors used data from 60 patients included in a previous study that occurred in the first 5 years of the described period.6

All patients had been recently diagnosed with NSCLC or there were grounds for a very strong suspicion. They were considered operable after physiological assessment and conventional staging procedures with or without whole body integrated PET/CT. Exams were carried out at different institutions, PET/CT was available from 2006. Tumor FDG uptake was qualitative and assessed by comparison to the background activity by visual inspection. Staging was based on the TNM 6th edition.5

All patients underwent thoracotomy and lung resection plus systematic lymphadenectomy. When exploratory thoracotomy revealed an unresectable tumor these patients were excluded.

Results were presented as a percentage for categorical and mean±standard deviation for continuous variables. Cohen's kappa coefficient was used to evaluate the agreement between clinical and pathological staging: 0.0, poor; 0.0–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; 0.81–1, almost perfect.16 Chi square and Student's t-tests were used to compare proportions and continuous variables, respectively. Statistical analysis was done using STATA 9.0 (Statistical/Data Analysis, Stata Corp., College Station, TX 77845, USA). Statistical significance was set to a value of p≤0.05.

This study allowed us to confirm that the inclusion of PET/CT in the preoperative assessment of NSCLC improved clinical staging, giving increased accuracy and the achievement of a good agreement with the pathological staging.

The agreement between clinical and pathological staging in the whole sample of 150 patients was 51.3%. This was a great improvement compared with the agreement obtained in the first study, which was only 21.7% and PET/CT was not available.6 Besides this progress, the overall agreement remains lower than desirable, with upstaging being the main reason for staging modifications.

With PET/CT, upstaging and FN became less frequent than in the subgroup that did not perform this exam (16.4% vs 55.1%) and downstaging and FP more common (16.4% vs 4.5%).

According to the literature FN, in the majority of cases, occurred with small tumors (<7mm in diameter) or tumors with relatively low metabolic activities.17,18 In this study, FN occurred due to a lymph node underestimation in 90.0%, namely cN0 to pN1 in 66.7%, which in practice did not reflect a need for a change in the treatment options. Moreover, adenocarcinoma with BA component was present in 50.0% of all cases, which may reflect the most indolent biological behaviour of this histological subtype.17,18 FP, are mainly documented when there is metabolical activity of infectious or inflammatory lesions.19 The authors found FP occurring after induction chemotherapy and restaging PET/CT in 60.0%. Literature data demonstrated that, although PET/CT has an important role in restaging, its accuracy is lower, and both FP and FN are more frequent.13 Percentage of decrease in the maximum standardized uptake value (SUV) in tumor and lymph nodes may be useful as a predictor of therapy response.12 In addition, the accuracy of PET/CT seems to be higher when it is repeated within 4–12 weeks after induction therapy and close to surgery.12 In our study, FP results may reflect both the unavailability of SUV measuring and the length of time between restaging PET/CT and surgery (30.2 (±15.1) days).

Finally, in 6 patients with PET/CT the positive result in cN2 was not confirmed, which is a limitation to point out. In the meantime, non-surgical approaches, such as EBUS-TBNA and transesophageal endoscopic ultrasound-guided fine needle aspiration (EUS-FNA), became available, although for only one of the included patients. Furthermore, some studies indicate that the combination of EBUS and EUS enable more complete access to the mediastinum, and, in the future, may be considered the gold standard for mediastinal staging.20,21 Additional studies are evaluating the impact of these techniques in the accuracy of staging.

The authors have no conflicts of interest to declare.