Design paper
Toward Less Futile Surgery in Non-Small Cell Lung Cancer? A Randomized Clinical Trial to Evaluate the Cost-Effectiveness of Positron Emission Tomography

https://doi.org/10.1016/S0197-2456(00)00119-7Get rights and content

Abstract

Non-small cell lung cancer can be cured if the patient is medically operable and the tumor resectable. Current diagnostic strategies are aimed to detect tumor deposits that preclude resection with curative intent. However, these strategies are rather inefficient, resulting in a large number of futile invasive procedures. In the early 1990s positron emission tomography (PET) showed promising results at its introduction in the clinic, especially in oncology. A large number of accuracy studies have reported that PET is superior to conventional imaging. However, whether PET ultimately improves patient outcome should ideally be assessed by means of a randomized controlled trial. No such design has been applied to evaluate PET in oncology so far. The PLUS study was designed to compare the current strategy of conventional methods with a strategy where PET was added after completion of noninvasive techniques. Patients considered operable by the physician at this point were then randomly assigned to PET and further consequences or to standard procedures of mediastinoscopy or thoracotomy. Primary outcome events were futile thoracotomies. The trial randomized 188 patients from nine hospitals in 1 year. Patient enrollment has been stopped and data collection is in progress. The results will be published in 2001. Control Clin Trials 2001;22:89–98

Introduction

A patient with non-small cell lung cancer (NSCLC) can be cured only if the patient is medically operable and the tumor resectable. The clinical diagnostic workup of such patients aims to establish a diagnosis of NSCLC (the primary tumor) preoperatively and to identify lymphatic or hematogeneous metastases to restrict curative surgery to patients who potentially benefit from this treatment. For this purpose, a battery of diagnostic tests examining the various target tissues [lymph nodes with chest computed tomography (CT) scan, bone with plain X-rays, skeletal scintigraphy, brain with CT/magnetic resonance imaging (MRI); liver and adrenals with CT or ultrasound] are at the clinician's disposal. Various guidelines have been formulated to optimize these efforts 1, 2, 3, 4. In daily clinical practice, however, variation is considerable 5, 6, 7, 8. In some hospitals, patients without enlarged mediastinal lymph nodes (MLN) at CT scanning proceed directly to thoracotomy. In other hospitals, mediastinoscopy is part of the standard workup. Mediastinoscopy is always performed in patients with enlarged N2 MLN. Sensitivity and specificity of CT in detecting mediastinal nodes are, respectively, 52% and 69% so that patients with negative CT who proceed to thoracotomy may in fact have N2 MLN disease [9]. Mediastinoscopy itself may also be inaccurate. Thus, the prevailing diagnostic strategy leads to futile thoracotomies in up to 50% of patients 8, 10, 11.

In the early 1990s positron emission tomography (PET) emerged as a promising diagnostic imaging tool in nuclear medicine. In the oncological setting, the combination of whole-body PET with the tumor-seeking radioactive tracer fluorodeoxyglucose (18FDG) allows for noninvasive visualization and quantification of tumor deposits. Since PET provides spatial and metabolic information, it can assume a complementary role to currently available morphological imaging, or even surpass these techniques. A large number of published accuracy studies have reported, without exception, that PET is superior to the CT scan in assessing the nodal, i.e., mediastinal tumor status [12]. However, accuracy studies are not only highly susceptible to bias, but also fail to cope with the full complexities of how an imaging technique contributes to the diagnostic and therapeutic process 13, 14, 15, 16, 17. By definition, the design requires that the technique be assessed out of the clinical context [18]. More accurate staging may lead to more appropriate treatment planning, but prospective evidence for that is largely anecdotal [19]. Such evidence is subsequently used in decision analyses, where risks, benefits, and costs of various strategies are modeled mathematically [20]. Several of these studies suggest that PET is cost-effective in NSCLC staging 21, 22. Given these limitations, ideally a randomized controlled trial (RCT) should assess whether PET provides information that ultimately improves patient outcome 18, 20, 23.

Where patients are selected for curative surgery, the achievable health gain consists of a reduction of the number of futile surgical procedures. The PLUS study (PET in LUng cancer Staging) was designed to compare the current strategy of conventional methods with a strategy where PET was added to the noninvasive techniques with respect to these outcomes. Since the introduction of new health technology for specific clinical indications requires not only justification in terms of effectiveness but also in terms of costs, data on direct medical costs were collected concurrently.

Section snippets

Population

Eight community hospitals and one academic hospital recruited patients for the trial. Patients with suspected or proven NSCLC, considered medically operable and potentially resectable by the local pulmonary physician on the basis of clinical staging procedures (i.e., clinical stage I–III), but prior to surgical staging, were invited to participate (see Figure 1). The diagnostic routine in each hospital was respected. Inclusion and exclusion criteria were minimal to guarantee fast accrual and

Discussion

The introduction of new health technology requires justification of cost-effectiveness for specific clinical indications. In the case of a diagnostic imaging technique, effectiveness means that it not only provides more accurate data than existing modalities, but also improves patient management and ultimately has a favorable impact on health status at reasonable costs. In the development of new therapeutic agents, acceptance and reimbursement can only be achieved by at least two randomized

References (37)

  • Anonymous

    Pretreatment evaluation of non-small-cell lung cancer

    Am J Respit Crit Care Med

    (1997)
  • R.J.A.U. Fergusson et al.

    Management of lung cancer in South East Scotland

    Thorax

    (1996)
  • G.J.M. Herder et al.

    Staging non-small cell lung cancer patients in two Dutch hospitals

    Eur J Resp Dis

    (1999)
  • W.R. Webb et al.

    CT and MR imaging in staging non-small cell bronchogenic carcinomaReport of the Radiologic Diagnostic Oncology Group

    Radiology

    (1991)
  • P. Goldstraw

    The practice of cardiothoracic surgeons in the perioperative staging of non-small cell lung cancer

    Thorax

    (1992)
  • Investigation for mediastinal disease in patients with apparently operable lung cancer

    Ann Thorac Surg

    (1995)
  • B.A. Dwamena et al.

    Metastases from non-small cell lung cancerMediastinal staging in the 1990s—Meta-analytc comparison of PET and CT

    Radiology

    (1999)
  • C.B. Begg

    Biases in the assessment of diagnostic tests

    Stat Med

    (1987)
  • Cited by (0)

    *

    Participating pulmonologists of the PLUS study are: J.H.A.M. van den Bergh, Medical Centre Alkmaar; R.A.L.M. Stallaert, Westfries Gasthuis, Hoorn; A.J.M. Schreurs, Onze Lieve Vrouwe Gasthuis, Amsterdam; J.P. Teengs, Kennemer Gasthuis, Haarlem; J. Berkovits, Ziekenhuis Amstelveen; W.F. Strankinga, BovenIJ Ziekenhuis, Amsterdam; G. Visschers, Slotervaart Ziekenhuis, Amsterdam; and C. Jie, Lucas/Andreas Ziekenhuis, Amsterdam.

    View full text