| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2002;73:394-402
© 2002 The Society of Thoracic Surgeons
a Division of Cardiothoracic Surgery, College of Public Health, University of Iowa, Iowa City, Iowa, USA
b Division of Nuclear Medicine, College of Public Health, University of Iowa, Iowa City, Iowa, USA
c Department of Biostatistics, College of Public Health, University of Iowa, Iowa City, Iowa, USA
d Iowa City Veterans Administration Medical Center, Iowa City, Iowa, USA
* Address reprint requests to Dr Kernstine, Division of Cardiothoracic Surgery, University of Iowa Hospitals and Clinics, 200 Hawkins Dr, Room 1616 JCP, Iowa City, IA, USA 52242-1038
e-mail: kemp-kernstine{at}uiowa.edu
Presented at the Thirty-seventh Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 29–31, 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Methods. We performed a 5-year retrospective analysis of operable patients with known or suspicious NSCLC. Standard PET techniques were used. Inclusion criteria were (1) surgical mediastinal nodal sampling by mediastinoscopy within 31 days of the PET and (2) definitive diagnosis.
Results. There were 237 patients who met the evaluation criteria; ninety-nine patients with NSCLC and 138 with suspicious lesions (137 men and 100 women; aged 20 to 88 years). The PETs were performed from 0 to 29 days before mediastinoscopy (median, 7 days). The standardized uptake value for the primary lesion was 0 to 24.6 (7.9 ± 5.0). Nine primary lesions had no FDG uptake (1 benign, 8 NSCLCs). Seventy-one patients (31%) had mediastinal PET positive disease, and 44 patients (19%) had histologic positive mediastinal disease; N2 41 patients (17%) and N3 9 patients (4%). In 6 patients (3%), the initial frozen sections were negative, but PET positivity encouraged further biopsies that were positive for cancer. The PET sensitivity was 82%, specificity 82%, accuracy 82%, negative predictive value 95%, and positive predictive value was 51%. All primary lesions with a standardized uptake value less than 2.5 and a negative mediastinal PET were negative histologically (n = 29). Logistic regression analysis resulted in 100% specificity for PET in this group.
Conclusions. In NSCLC PET may reduce the necessity for mediastinoscopy when the primary lesion standardized uptake value is less than 2.5 and the mediastinum is PET negative. Accepting this approach in our patient population, the need for mediastinoscopy would have been reduced by 12%.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
In NSCLC patients, clinical trials of computed tomography, magnetic resonance imaging, and 2-(18F) fluoro-2-deoxy-D-glucose positron emission tomography (PET) have been performed to determine whether the mediastinal lymph nodes (MLN) are involved with metastatic NSCLC [3–5]. The PET appears to be particularly promising in the following areas:
2. In 10% to 12% of potentially operable patients PET will identify biopsy-proven metastasis, not identified by other tests [8–10].
3. The PET can determine the metabolic activity of the primary lesion, which may be an important, independent prognostic indicator [11].
Few PET research studies have had sufficient patients, particularly those with MLN involvement, to determine whether PET, by itself, may be used to stage the mediastinum. Furthermore, many prior clinical reports include fine needle aspiration of the primary or MLN, or in some cases no biopsies at all, thus further decreasing our confidence in PET as a means to stage the mediastinum.
Realizing these deficiencies in the previously reported trials, we embarked on an analysis of data acquired from our Thoracic Surgical Clinic at The University of Iowa intended to compare PET with mediastinal surgical staging by mediastinoscopy. Our goal was to determine if PET could reduce or eliminate the need for mediastinal surgical biopsy in patients with known or suspicious NSCLCs.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Radiologic imaging
Positron emission tomography: 2-(18F) fluoro-2-deoxy-D-glucose (F-18 FDG) was produced by a Nuclear Interface FDG synthesizer using a modified (Hamacher, Uppsala, Sweden) nucleophilic fluorination method [12, 13]. Patients had to fast for 6 hours and the PET was performed if serum glucose was less than 200 mg/dL [14]. Each patient was injected with 10 mCi (370 MBq) of F-18 FDG. After an uptake period of 60 minutes, the patients were imaged using a GE 4096 Plus PET whole body PET scanner (GE Medical Systems, Milwaukee, WI) from the base of the brain to the level of the proximal thighs in 8 to 10 bed positions. The system produces 15 transverse tomographic slices over a 10 cm axial field of view and has a spatial resolution of 6.5 mm full-width-half-maximum. Transmission scans using a rotating Ge-68 pin source were performed on all patients after the acquisition of the F-18 FDG data to correct for attenuation. The corrected emission data were reconstructed with the filtered backprojection algorithm using a 10-mm Hanning filter. Coronal, sagittal, and transverse plane tomographic images were generated with and without attentuation correction and viewed on computer monitor.
Calculation of standardized uptake values
The standardized uptake value (SUV) represents the amount of radioactivity in the area of interest adjusted by the patient’s weight and dose of FDG injected corrected for the radioactive tracer decay:
 |
The mediastinum was visually assessed. Regions that appeared to be significantly greater than mediastinal background activity were considered positive for metastatic disease. The nodal regions were then grouped as ipsilateral (N2 nodes) or as contralateral (N3 nodes).
Surgical staging
Surgeons were aware of the PET test results. Cervical and anterior mediastinoscopy were performed for a histologic sampling only. Anterior mediastinoscopy was performed when the primary lesion originated in the left upper lobe or left hilum, or when there was suspicious adenopathy in the anterior mediastinum or aortopulmonary window.
Statistical analysis
The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated for PET tests using the surgical MLN biopsy results as the standard. When the primary lung lesion was not visible in the PET images, it was assigned an SUV of zero. For analysis, the American Thoracic Society nodal stations that were biopsied were grouped into N2 and N3 depending upon the side of the primary lesion. Thus, each patient would have no more than 2 MLN groups entered for the analysis, independent of the number of biopsies taken. Data were recorded for statistical analyses in Microsoft EXCEL spreadsheets (Microsoft Corp, Redmond, WA). In addition to descriptive statistics (eg, means and percentages) obtained using the EXCEL functions, several SAS procedures were used to perform analyses (SAS software, SAS Institute Inc, Cary, NC). These SAS procedures included the FREQ (for Fischer’s exact test, for kappa statistics, and for exact 95% confidence intervals on sensitivity, specificity, and predictive values), the TTEST (for t tests), the NPAR1WAY (for Wilcoxon tests), and the LOGISTIC (for logistic regression) procedures. The alpha level for statistical significance was retained at p = 0.05. Reported confidence intervals (CIs) are 95%. Distribution widths are reported as ± standard deviation.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
The characteristics of the study group may be found in Table 1. Patients were grouped into those that were known at presentation to have NSCLC (99 patients) and those patients that were suspicious for NSCLC (138 patients). The two groups were similar in characteristics by age (t test, p = 0.24; Wilcoxon test, p = 0.25) and gender (Fisher’s exact test, p = 0.59). There were 137 males and 100 females, with a mean age of 65 ± 11 years ranging from 20 to 88 years. There appeared to be more right-sided tumors in the suspect group (Fisher’s exact test, p = 0.0002).
|
There were 138 patients in the suspected NSCLC group (suspect group). The mean age of the patients was 66 ± 11 years, and there were 82 men and 56 women. Ninety-three patients (67%) had right-sided primaries. The PET scan was performed 9 ± 8 days (range, 0 to 29 days) before mediastinoscopy. Seventeen scans (12%) were performed more than 3 weeks before mediastinoscopy. The SUV of the primary was 6.6 ± 4.7 (0 to 20.5). Thirty-six patients (26%) had mediastinal disease noted on PET scan. Eighteen patients (13%) had histologic positive results by mediastinoscopy.
The primary tumor cell-type was adenocarcinoma in 44 patients (32%), bronchioalveolar in 5 (4%), squamous cell in 26 (19%), large cell in 1 (1%), nonspecified NSCLC in 19 (14%), carcinoid in 4 (3%), atypical carcinoid in 1 (1%), metastatic carcinomas in 3 (3%); 2 leiomyosarcomas and 1 transitional cell sarcoma (3%), small cell lung cancer in 5 (4%), and 31 benign lesions; granulomas in 16 (12%), inflammation in 1 (1%), fibrosis in 4 (3%), and others in 10 (7%).
Table 1 also lists the combined group, those patients with known NSCLC before their mediastinoscopy and those who had suspected NSCLC. The mean age for the entire group is 65 ± 11 with a median of 68 years and a range of 20 to 88 years, and the midquartile range is from 57 to 74 years (ie, half the patients were in this age range). There were 137 men and 100 women. One hundred and thirty five of the cases were right-sided lesions (56%). The PET was performed 9 ± 8 days before the mediastinoscopy with a range of 0 to 29 days, and the quartile range is from 2 to 14 days. In 26 of the patients (11%) the PET was performed more than 3 weeks before mediastinoscopy. There were 71 patients (31%) that had mediastinal disease on PET. Forty-four patients (19%) had a histologic positive disease by mediastinoscopy. The final tally of both groups of the tumor cell types can be found in Table 1.
It appeared there was a difference for the days between PET and mediastinoscopy for the known group and the suspect group. The patients in the known group appeared to wait an average 7 days before their mediastinoscopy, whereas the patients in the suspect group appeared to wait approximately 9 days; this appeared to be statistically greater in the suspect group (t test, p = 0.0445; Wilcoxon test, p = 0.0263). However, the number of patients waiting longer than 3 weeks between the 2 groups was not statistically different (Fisher’s exact test, p = 0.53).
The primary SUVs in the known group appeared to be higher than the SUVs in the suspect group. The mean primary SUV in the known group was 9.5 and in the suspect group, 6.6. Not surprisingly, the inclusion of the benign lesions reduced the SUV in the suspect group (t test, p < 0.0001; Wilcoxon test, p < 0.0001).
Also, there appeared to be more patients found to have mediastinal metastases in the NSCLC group (Fisher’s exact test; p = 0.0114). There were 26 of 99 patients (26%) in the known group and 18 of 138 patients (13%) in the suspect group. Again, this is not a surprising result since the benign patients are included in the suspect group. The PET identified a higher metastatic cancer rate in both groups: 35% in the known group and 36% in the suspect group, with 30% to 50% more than would be found by mediastinoscopy. When comparing the two groups for the nine cancer cell-type categories (adenocarcinoma, bronchioalveolar, squamous cell, large cell, nonspecified NSCLC, carcinoid, atypical carcinoid, metastatic cancer, and small cell), there did not appear to be any differences (Fisher’s exact test; p = 0.0718).
Table 2 represents the statistical analyses comparing the PET data to the final histologic data from mediastinoscopy. The table is separated into the 3 groups: (1) The known NSCLC group before mediastinoscopy (n = 99), (2) The suspect NSCLC group before mediastinoscopy (n = 138), and (3) The combined groups (known and suspect NSCLC) groups (n = 237).
|
In the suspect group there were 16 patients (12%) that were N2 histologically positive and 5 (4%) that were N3 positive. There are a total of 18 patients (13%) that were positive in either N2 or N3. Two patients that were N3 positive were also N2 negative. There were 22 patients whose primary PET SUV was less than 2.5 and the mediastinum was PET negative, and again none of these patients had histologic positive disease. Again, the logistic regression demonstrated 100% specificity.
In the combined (known and suspect) groups (n = 237), there were 41 patients (17%) that were N2 histologically positive and 9 (4%) that were N3 positive. The total, whether the patients had N2 or N3 positivity was 44 (19%), 3 of whom were N3 positive and N2 negative. There were 29 patients whose primary PET SUV was less than 2.5 and the mediastinum was PET negative, and none of those patients demonstrated histologic positive disease. The logistic analysis also demonstrated 100% specificity.
Sensitivity and specificity
N2 results
In the known group the sensitivity was 0.80 (CI, 0.59 to 0.93). The specificity also rounded to 0.80 (CI, 0.69 to 0.88). Specificity was estimated with more precision simply because there were more negative mediastinal results than positive results. Indeed, the proportion of positive N2 mediastinal results was 0.25 (CI, 0.17 to 0.35). The predictive value of a positive N2 PET result was 0.57 (CI, 0.39 to 0.74), and the predictive value of a negative N2 PET result was 0.92 (CI, 0.83 to 0.97).
In the suspect group the N2 sensitivity was 0.81 (CI, 0.54 to 0.96) and the N2 specificity also rounded to 0.81 (CI, 0.73 to 0.88). These results are equivalent to those obtained from the Known group. The proportion of positive N2 mediastinal results among these patients was 0.12 (CI, 0.07 to 0.18). The predictive value of a positive N2 PET result was 0.36 (CI, 0.21 to 0.54). The predictive value of a negative N2 PET result was 0.97 (CI, 0.92 to 0.99).
N3 results
In the known group the sensitivity was 0.25 (CI, 0.01 to 0.81), whereas the specificity was 0.96 (CI, 0.90 to 0.99). A positive N3 result is extremely rare and the expected probability of a positive mediastinal result is 0.04 (CI, 0.01 to 0.10). The predictive value of a positive N3 PET result is 0.20 (CI, 0.01 to 0.72). The predictive value of a negative N3 PET result is 0.97 (CI, 0.91 to 0.99).
In the suspect group the N3 sensitivity was 0.00 (CI, 0.00 to 0.52), whereas the N3 specificity was 0.99 (CI, 0.95 to 1.00). Again, the expected probability of a positive mediastinal result is 0.04 (CI, 0.01 to 0.08). The predictive value of a positive N3 PET result is 0.00 (CI, 0.00 to 0.84). The predictive value of a negative N3 PET result is 0.96 (CI, 0.92 to 0.99).
Combined N2 and N3 results
In the known group the overall sensitivity was 0.81 (CI, 0.61 to 0.93) and the overall specificity was 0.81 (CI, 0.70 to 0.89). The overall expected probability of a positive mediastinoscopic result is 0.26 (CI, 0.18 to 0.36). The predictive value of a positive PET result is 0.60 (CI, 0.42 to 0.76). The predictive value of a negative PET result is 0.92 (CI, 0.83 to 0.97).
In the suspect group the overall sensitivity is 0.83 (CI, 0.59 to 0.96) and the overall specificity is 0.83 (CI, 0.75 to 0.89). The expected probability of a positive mediastinoscopic result is 0.13 (CI, 0.08 to 0.20). The predictive value of a positive PET result is 0.42 (CI, 0.26 to 0.59). The predictive value of a negative PET result is 0.97 (CI, 0.92 to 0.99).
Combined group (known + suspect NSCLC) sensitivity and specificity
In the 237 combined group the N2 sensitivity was 0.80 (CI, 0.65 to 0.91). The N2 specificity was 0.81 (CI, 0.74 to 0.86). The proportion of positive N2 mediastinoscopic results was 0.18 (41 of 237), (CI, 0.13 to 0.23). The predictive value of a positive N2 PET result is 0.46 (CI, 0.35 to 0.59), and the predictive value of a negative N2 PET result is 0.95 (CI, 0.91 to 0.98). The observed N2 accuracy was 0.81.
For these 237 combined group patients, the N3 sensitivity was 0.11, (CI: 0.003 to 0.48). The N3 specificity was 0.97 (CI, 0.94 to 0.99). The proportion of positive N3 mediastinoscopic results was 0.04 (9 of 237), (CI: 0.02 to 0.07). The predictive value of a positive N3 PET result is 0.14 (CI, 0.004 to 0.58). The predictive value of a negative N3 PET result is 0.97 (CI, 0.93 to 0.98). The observed N3 accuracy was 0.94.
Finally, for these 237 combined group patients, the "any" (N2+N3 mediastinal node sensitivity was 0.82 (CI, 0.67 to 0.92), and the "any" node specificity was 0.82 (CI, 0.76 to 0.87). The proportion of positive N2 mediastinoscopic results is 0.19 (44 of 237) (CI, 0.14 to 0.24). The predictive value of a positive PET result is 0.51 (CI, 0.39 to 0.63), and the predictive value of a negative PET result is 0.95 (CI, 0.91 to 0.98). The observed "any" node accuracy is 0.82.
A small, but particularly important group of 6 patients (3%) (4 in the known group and 2 in the suspect group) who initially had negative mediastinal node biopsy by mediastinoscopy, but because of positive PET images were further biopsied and were found to be positive. Without the PET guidance these patients would have been under-staged and inappropriately treated. Positron emission tomography improved the sensitivity of mediastinoscopy to identify metastatic disease.
Positron emission tomography accuracy was similar in the two known and suspect groups. For the known group the observed "any" mediastinal accuracy was 0.808. The estimated kappa was 0.554 ± 0.089 (standard error). For the Suspect group the observed "any" mediastinal node accuracy was 0.826. The estimated kappa was 0.462 ± 0.88 (standard error). These kappa estimates do not differ between the two patient groups (p = 0.46).
Within the chosen 30-day time limitation, the timing of PET to the performance of mediastinoscopy did not appear to affect the statistics obtained in these results. The percentage of patients that had mediastinoscopy more than 3 weeks after PET was relatively few, 11% in both groups. Using a logistic regression analysis to examine the effect of the length of time between the PET and the mediastinoscopy, there did not appear to be any significant increase in the number of histologic positive nodal stations when the PET was negative (p = 0.70). The odds ratio (per daily increase) was 0.99 with a 95% CI from 0.94 to 1.05.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
This analysis represents a retrospective evaluation from a single tertiary institution. Our analysis is one of the largest reported trials evaluating the use of PET to stage the mediastinum. As a result of the large number of patients in this trial, we were able to analyze subsets (N2 and N3), in addition to the combination (N2 + N3). We also subdivided the patients into Known NSCLC and Suspect NSCLC groups because (1) it represented how patients clinically present to thoracic surgeons, and (2) given the prevalence of granulomatous disease in our treating region, the Midwestern United States, PET may be less helpful in the Suspect group with potentially more benign inflammatory disease. However, our statistical findings are the same in both groups (Table 2). The increased frequency of granulomatous disease in the Suspect group does not appear to significantly affect the ability of PET to stage the mediastinum.
It should be pointed out that we are really comparing the PET with mediastinoscopy plus PET, not with mediastinoscopy alone. Positron emission tomography scans are more commonly performed before thoracic surgery consultation. As a result of this paradigm shift, we have learned to work with the PET information to enhance our evaluation of the NSCLC patient. Positron emission tomography adds greatly to the workup and as a result our question, "Can PET reduce the need for mediastinoscopy...?" is more relevant to clinical practice. So we are not questioning the use of PET in the evaluation of NSCLC; instead, we are assuming that clinicians are now questioning the use of mediastinoscopy once they have performed a PET scan. We are anticipating health maintenance organizations and Hospital Consortium Funding Association determining that mediastinoscopy is unnecessary if a PET scan has been performed. So our question is: Does mediastinoscopy provide additional helpful clinical information to the PET scan? From this report we say that it does in 88% of the patients studied. To eliminate mediastinoscopy, PET should be at least as accurate as mediastinoscopy and it is not. Positron emission tomography did not appear to have sufficient positive predictive value and negative predictive value (Table 2). If the PET is negative in either the N2 or N3 nodal groups by mediastinoscopy, there is a 1% to 8% likelihood that there is mediastinal cancer; but when the N2 is PET positive, there is a 40% to 60% likelihood that there is no metastatic cancer. There are insufficient N3 positive nodes to make a statement about the positive predictive value for PET; the 95% confidence intervals were large (60% to 80%). So if the patient is PET positive in the N2 or N3 location, we advise that mediastinoscopy should be performed. In addition, PET assists mediastinoscopy, we found an extra 6% with mediastinal disease because of PET.
Although it does not eliminate it, PET does reduce the need for mediastinoscopy. We identified a group of patients in whom the primary lesion PET SUV was less than 2.5 and the mediastinum was PET negative (29 of 237) (Table 2). Because we did not include other important clinical information, we are reluctant to advise against mediastinoscopy in all of these patients. For example, in the case of hilar masses, which are more likely to spread to the mediastinum, metastatic disease may be missed. We believe that it is certainly appropriate to eliminate mediastinoscopy using the PET alone for peripheral lesions that have an SUV less than 2.5 and all mediastinal PETs negative. Further clinical information may provide greater direction.
To illustrate this concept, Figure 1 shows an example of a chest computed tomographic scan and FDG-PET image from a patient with a suspicious right lower lobe lesion. The lesion was peripheral and had no FDG uptake; thus the SUV was less than 2.5 and the mediastinum was PET negative. However, the lesion appeared suspicious on computed tomography. Mediastinoscopy demonstrated no histologic positive nodes and the patient underwent a right thoracoscopy confirming the diagnosis of NSCLC, and a right lower lobectomy was performed. This case illustrates two important issues when dealing with PET and NSCLC: (1) the inability of PET to definitively exclude malignant disease in the lung; and (2) the PET results would have appropriately directed us away from performing mediastinoscopy.
|
However, our goal was to see if mediastinoscopies could be satisfactorily reduced in our study group. We are not questioning the true rate of cancer in the mediastinal nodes, so we did not include the nodal results found at thoracotomy.
Also, our question is solely on the use of PET alone without any other information. Certainly it would be more appropriate to include other information such as the computed tomography, tumor size, border, location, nodal size, and so forth, but it is our impression that clinicians are more commonly using PET as the sole means to stage the mediastinum. From our analysis using PET alone to stage the mediastinum is incorrect. Careful review of all clinically relevant information should be performed.
In summary, we believe that PET provides helpful clinical information in patients with operable NSCLC or with a suspicious, possibly operable NSCLC (Fig 2). Positron emission tomography does not eliminate the need for mediastinoscopy. However, it will select a group in which mediastinoscopy may be unnecessary. Furthermore, it will assist in those patients that require surgical mediastinal biopsy to improve the accuracy of surgical mediastinal staging.
|
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
I think you have nicely shown that when the positron emission tomography (PET) is negative and the standardized uptake valves (SUVs) are low, maybe one can avoid mediastinoscopy. But the much more important message that we need to get out to the nonsurgeons is that when the PET is positive, patients should not get enrolled in nonsmall cell lung cancer (NSCLC) protocols and miss seeing the surgeon for mediastinoscopy. In Birmingham, Alabama, we have a high rate of histoplasmosis. We are seeing a lot of false-positive PETS. We are in the middle of a cost-effective analysis of PET. It calculates the cost of unnecessary procedures (ie, Chamberlains, mediastinoscopies, biopsies of other lesions), and I wonder if you have looked at the cost of the PET itself and then the cost of tests or procedures it adds that are negative or positive or both. We are often chasing our tail going after these PET lesions, and if we had not ordered the PET scan, we would have just operated on the patient like you end up doing in the long run. In our experience, this both downstages and upstages patients.
I also wonder if you could talk a little bit about the learning curve for some of the places that are just starting to develop a new PET scan center. We have found there is a learning curve of reading a PET. There is really no anatomy and a lot of the physicians and radiologists who are reading it are not really nuclear radiologists and really do not know chest anatomy well. They can not tell you an N1 from an N2. And the final issue is, I saw that you skipped over N1 nodes, and I am wondering if you are doing that intentionally, because we have found that large Ts obscure the N1 and PET can not give you any information on the N1. Overall, we believe PET is useful, but when PET is positive, tissue biopsies must be obtained. I appreciate your work and look forward to your comments on these questions.
DR KERNSTINE: Concerning the issue of PETs, I guess you are referring to the positive predictive value. For the group overall the positive predictive value is 51%, and the range is approximately 40% to 80%. In essence, when PET says there is a positive node, it is incorrect in half. In my mind, thus, as Dr Cerfolio is saying, we would still need to do a mediastinoscopy in that patient population. Then, on the converse, when PET is negative, it is incorrect between 2% and 8%, the 95% confidence interval. Again, I do not feel that that is accurate enough. Considering the safety of mediastinoscopy, I would prefer to have something say zero or 1% to deny a patient further evaluation.
Now, I think your next question was concerning the experience of the radiologist. We found the same problem in Iowa. There are four new PET scanners in our region, and there are many questions from medical oncologists. For those that are not calling, patients are being treated with induction therapy based on no histologic sampling, rather, the PET results alone. This is one of the reasons we performed this study. These other programs are not performing histologic or cytologic evaluation of potential metastatic lesions.
Concerning your question about the N1 nodes, we excluded N1 evaluation because N1 status did not appear to be of any help in answering our question: the role of PET in determining the need for mediastinoscopy. I think, however, that it is something we should evaluate in another study. We do collect all that information, but we were more interested in what its role was for mediastinoscopy.
DR JAMES D. LUKETICH (Pittsburgh, PA): Very nice presentation. I think you are very wise to try to stick with one very specific question with PET. It gets very confusing when you try to answer them all with one study.
Your patients who had SUVs less than 2.5, were these all very small lesions? As you get to a lesion that is less than a centimeter on computed tomography, I think you can predict that you are going to miss some of those by PET even if they are malignant.
Nasser Altorki summarized the data very nicely the other day on small pulmonary lesions and noted that up to 10% have microscopic mediastinal nodal mets. Do you have any suggestion for what to do about that relatively small but not insignificant group?
In addition to that, are you doing any combined molecular studies? Obviously, we don’t know the clinical impact of this, but there is clearly a significant number of patients who are recurring who have stage 1 disease with a negative histologic mediastinum. Are you doing any companion molecular studies?
Finally, I wonder if you could mention anything about some of the newer markers or the newer tracers, like C-11 choline. There are a couple of very interesting reports on C-11 choline that might drive the accuracy of lymph node staging in some reports as high as 99%. I wonder if you have any experience or if you have had the opportunity to review any of that data.
DR KERNSTINE: Thank you, Dr Luketich. When it comes to PET, size does matter. As far as the primary lesion, the larger the lesion, the greater the sensitivity, the specificity, and the accuracy. We were specifically dealing with PET and its use to evaluate the mediastinum. We did not evaluate the use of computed tomographic scanning in this patient population. In future analyses we are going to need to take all those factors into consideration. In this study, we are emphasizing the point that you can not use a single test to evaluate patients for potential metastatic disease.
As far as companion studies, we are involved in CALGB 9761, the NSCLC marker analysis 1 trial. To date, we have not initiated any other marker trials. I agree that these may be important to evaluate patients.
Concerning C-11 choline, we do have a cyclotron at the University of Iowa, and we have made C-11 and have injected patients. Most of the evaluations have been prostate cancer, not lung cancer. There is a variety of other tracers useful for not only evaluating the likelihood of malignancy but also how the tumor responds or might respond to chemotherapy. For example, there are some water markers that might help us determine the vascularity of tumors and determine the potential that an anti-angiogenesis chemotherapeutic agent may be helpful for a given tumor.
DR ERIC VALLIERES (Seattle, WA): Nice work, Kemp, once again. Your group has been a leader in the field of PET scanning for lung cancer in our country and the work you produce is always of good quality.
I would echo James Luketich’s comments about the relative importance of the size of imaged primary lesion when addressing a PET question. My feeling is that of the 12% of the lesions in whom you would have spared a mediastinoscopy by using the guidelines defined, most lesions were probably small, peripheral and some may have been of BAC histology. I think that it would be worthwhile going back to your data and addressing this. Also, for the purpose of this study, you eliminated patients with N1 disease identified by PET. Possibly if you had included these clinical N1 patients, the incidence of false negative mediastinal PETs would have been higher. Perhaps you may want to undertake another review of your data in the future. This is good work and I would encourage you to take it a little further. Thank you.
DR KERNSTINE: Thank you.
DR JONATHAN E. RHOADS (York, PA): If I understood what you said, you said that you did not trust a negative PET scan, that the lymph nodes in the mediastinum would be negative, because of approximately a 6% false-negative rate, and you said that if the lymph nodes in the mediastinum were positive, it is a false-positive rate in about 50%. The PET scan is very expensive. Why bother to do the scan?
DR KERNSTINE: As I pointed out in the algorithm, approximately half of the 15% to 20% found to have metastases by PET, or about 8% to 10%, will be confirmed to have metastatic disease. These patients would have undergone surgical treatment unnecessarily. Second, there is, of course, a group of patients that we have identified that we would not need to do mediastinoscopy. We could go directly to lung resection. And, of course, mediastinoscopy has some cost to it. It is going to take some time for the people to do the frozen section, for you to get in and do your biopsy, and then there is some risk associated with it as well, albeit very small. Finally, PET appears to improve the accuracy of mediastinoscopy by increasing sensitivity.
DR WALTER J. SCOTT (Omaha, NE): Nice article. Perhaps I missed it, but what is your definition of a positive mediastinum on PET? Is that based on a SUV or some other scale? I appreciate your work. Thank you.
DR KERNSTINE: And that is a concern for me as well, the lack of objectivity determining whether most institutions, including ours, perform a subjective assessment; and this is where having an experienced PET reader comes into play. The mediastinal PET image is reviewed to determine whether there are distinct areas above background. No actual number, such as SUV, can reliably be placed on it.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. K. Sit, A. D. Sihoe, W. S. Suen, and L. C. Cheng Positron-Emission Tomography for Lung Cancer in a Tuberculosis-Endemic Region Asian Cardiovascular and Thoracic Annals, February 1, 2010; 18(1): 33 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Tasci, C. Tezel, A. Orki, O. Akin, O. Falay, and C. A. Kutlu The role of integrated positron emission tomography and computed tomography in the assessment of nodal spread in cases with non-small cell lung cancer Interact CardioVasc Thorac Surg, February 1, 2010; 10(2): 200 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Stiles, E. L. Servais, P. C. Lee, J. L. Port, S. Paul, and N. K. Altorki Point: Clinical stage IA non-small cell lung cancer determined by computed tomography and positron emission tomography is frequently not pathologic IA non-small cell lung cancer: the problem of understaging. J. Thorac. Cardiovasc. Surg., January 1, 2009; 137(1): 13 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Morikawa, Y. Demura, T. Ishizaki, S. Ameshima, I. Miyamori, M. Sasaki, T. Tsuchida, H. Kimura, Y. Fujibayashi, and H. Okazawa The Effectiveness of 18F-FDG PET/CT Combined with STIR MRI for Diagnosing Nodal Involvement in the Thorax J. Nucl. Med., January 1, 2009; 50(1): 81 - 87. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Uesaka, Y. Demura, T. Ishizaki, S. Ameshima, I. Miyamori, M. Sasaki, Y. Fujibayashi, and H. Okazawa Evaluation of Dual-Time-Point 18F-FDG PET for Staging in Patients with Lung Cancer J. Nucl. Med., October 1, 2008; 49(10): 1606 - 1612. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N FUWA, T DAIMON, T MITSUDOMI, Y YATABE, T KODAIRA, H TACHIBANA, T NAKAMURA, T KATO, and Y SATO Identifying patients with peripheral-type early non-small cell lung cancer (T1N0M0) for whom irradiation of the primary focus alone could lead to successful treatment Br. J. Radiol., October 1, 2008; 81(970): 815 - 820. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Bordoni Consensus Conference: Multimodality Management of Early- and Intermediate-Stage Non-Small Cell Lung Cancer Oncologist, September 1, 2008; 13(9): 945 - 953. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Melek, M. Z. Gunluoglu, A. Demir, H. Akin, A. Olcmen, and S. I. Dincer Role of positron emission tomography in mediastinal lymphatic staging of non-small cell lung cancer Eur J Cardiothorac Surg, February 1, 2008; 33(2): 294 - 299. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. D. Vincent, E. El-Bayoumi, B. Hoffman, P. Doelken, J. DeRosimo, C. Reed, and G. A. Silvestri Real-Time Endobronchial Ultrasound-Guided Transbronchial Lymph Node Aspiration Ann. Thorac. Surg., January 1, 2008; 85(1): 224 - 230. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. C. Ung, D. E. Maziak, J. A. Vanderveen, C. A. Smith, K. Gulenchyn, C. Lacchetti, W. K. Evans, and Lung Cancer Disease Site Group of Cancer Care Onta 18Fluorodeoxyglucose Positron Emission Tomography in the Diagnosis and Staging of Lung Cancer: A Systematic Review J Natl Cancer Inst, December 5, 2007; 99(23): 1753 - 1767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C. Lee, J. L. Port, R. J. Korst, Y. Liss, D. N. Meherally, and N. K. Altorki Risk Factors for Occult Mediastinal Metastases in Clinical Stage I Non-Small Cell Lung Cancer Ann. Thorac. Surg., July 1, 2007; 84(1): 177 - 181. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Gregoire, K. Haustermans, X. Geets, S. Roels, and M. Lonneux PET-Based Treatment Planning in Radiotherapy: A New Standard? J. Nucl. Med., January 1, 2007; 48(1_suppl): 68S - 77S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Schimmer, K. Neukam, and O. Elert Staging of non-small cell lung cancer: clinical value of positron emission tomography and mediastinoscopy Interact CardioVasc Thorac Surg, August 1, 2006; 5(4): 418 - 423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Lardinois New Horizons in Staging for Non-Small-Cell Lung Cancer J. Clin. Oncol., April 20, 2006; 24(12): 1785 - 1787. [Full Text] [PDF]
|
|
|
|
|
|
J. L. Port, R. S. Andrade, M. A. Levin, R. J. Korst, P. C. Lee, D. E. Becker, and N. K. Altorki Positron emission tomographic scanning in the diagnosis and staging of non-small cell lung cancer 2 cm in size or less J. Thorac. Cardiovasc. Surg., December 1, 2005; 130(6): 1611 - 1615. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Matsumoto, M. Oda, Y. Tsunezuka, M. Tamura, K. Kawakami, and G. Watanabe Experimental study of extracorporeal lung resection in dogs: Ex situ sleeve resection and autotransplantation of the pulmonary lobe after extended pneumonectomy for central lung cancer J. Thorac. Cardiovasc. Surg., May 1, 2004; 127(5): 1343 - 1349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Inoue, N. Sawabata, S.-i. Takeda, M. Ohta, Y. Ohno, and H. Maeda Results of surgical intervention for p-stage IIIA (N2) non-small cell lung cancer: Acceptable prognosis predicted by complete resection in patients with single N2 disease with primary tumor in the upper lobe J. Thorac. Cardiovasc. Surg., April 1, 2004; 127(4): 1100 - 1106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. G. Pfister, D. H. Johnson, C. G. Azzoli, W. Sause, T. J. Smith, S. Baker Jr, J. Olak, D. Stover, J. R. Strawn, A. T. Turrisi, et al. American Society of Clinical Oncology Treatment of Unresectable Non-Small-Cell Lung Cancer Guideline: Update 2003 J. Clin. Oncol., January 15, 2004; 22(2): 330 - 353. [Full Text] [PDF]
|
|
|
|
|
|
J. L. Port, M. S. Kent, R. J. Korst, R. Keresztes, M. A. Levin, and N. K. Altorki Positron emission tomography scanning poorly predicts response to preoperative chemotherapy in non-small cell lung cancer Ann. Thorac. Surg., January 1, 2004; 77(1): 254 - 259. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. H. Kernstine Positron emission tomography with 2-[18F]fluoro-2-deoxy-D-glucose: can it be used to accurately stage the mediastinum in non-small cell lung cancer as an alternative to mediastinoscopy? J. Thorac. Cardiovasc. Surg., December 1, 2003; 126(6): 1700 - 1703. [Full Text] [PDF]
|
|
|
|
|
|
G. V. Gonzalez-Stawinski, A. Lemaire, F. Merchant, E. O'Halloran, R. E. Coleman, D. H. Harpole, and T. A. D'Amico A comparative analysis of positron emission tomography and mediastinoscopy in staging non-small cell lung cancer J. Thorac. Cardiovasc. Surg., December 1, 2003; 126(6): 1900 - 1904. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. E. Reed, D. H. Harpole, K. E. Posther, S. L. Woolson, R. J. Downey, B. F. Meyers, R. T. Heelan, H. A. MacApinlac, S.-H. Jung, G. A. Silvestri, et al. Results of the American College of Surgeons Oncology Group Z0050 trial: the utility of positron emission tomography in staging potentially operable non-small cell lung cancer J. Thorac. Cardiovasc. Surg., December 1, 2003; 126(6): 1943 - 1951. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Vesselle, E. Turcotte, L. Wiens, and D. Haynor Application of a Neural Network to Improve Nodal Staging Accuracy with 18F-FDG PET in Non-Small Cell Lung Cancer J. Nucl. Med., December 1, 2003; 44(12): 1918 - 1926. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kimura, N. Iwai, S. Ando, K. Kakizawa, N. Yamamoto, H. Hoshino, and T. Anayama A prospective study of indications for mediastinoscopy in lung cancer with CT findings, tumor size, and tumor markers Ann. Thorac. Surg., June 1, 2003; 75(6): 1734 - 1739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Lardinois, T. Brack, A. Gaspert, T. Spahr, D. Schneiter, H. C. Steinert, and W. Weder Bronchoscopic radioisotope injection for sentinel lymph-node mapping in potentially resectable non-small-cell lung cancer Eur J Cardiothorac Surg, May 1, 2003; 23(5): 824 - 827. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Lardinois, A. Schallberger, D. Betticher, and H.-B. Ris Postinduction video-mediastinoscopy is as accurate and safe as video-mediastinoscopy in patients without pretreatment for potentially operable non-small cell lung cancer Ann. Thorac. Surg., April 1, 2003; 75(4): 1102 - 1106. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
