Lack of response to chemotherapy is a major problem in cancer treatment. By using positron emission tomography (PET) to assess tumor response to first-line therapy, oncologists can determine whether it's time to pursue another treatment regimen with curative potential.

Specifically, assessing response may be useful in two possible situations: to evaluate tumor response at the end of a full course of treatment, or to predict tumor response early in the course of a prolonged treatment regimen. In the first instance, early detection of treatment failure may permit a physician to institute a second-line therapeutic approach. In the second instance, accurately predicting treatment failure may allow the physician to substitute an alternative regimen, without subjecting the patient to the toxicity of the full course.

Advantages and limitations
The greatest advantage of metabolic FDG PET over CT or MRI is evident when assessing treatment effects, since the anatomic changes that follow treatment frequently lag behind tumor response in terms of viability. A mass of necrotic tumor/scar tissue may remain apparent on CT images for months after successful tumor eradication.¹ In this situation, the advantage that PET offers because of its ability to demonstrate presence or absence of tumor metabolism of FDG is even greater than in pretreatment tumor evaluation.

However, PET with FDG has its own limitations, in assessing tumor response, since FDG metabolism is seen in all cells and is not specific to viable tumor. Macrophages and leukocytes may densely infiltrate post-treatment masses of necrotic tumor, and FDG uptake resulting from this benign cellular response may be confused with viable tumor activity. To minimize the effect of this nonspecificity, clinical studies are needed to determine optimal quantitative thresholds of FDG uptake for separating benign uptake from tumor activity.

Measuring change in FDG metabolism
In assessing end response, a single PET study can determine presence or absence of increased metabolic activity at the tumor site. For predicting response, quantitative or semiquantitative comparison of pretreatment and early treatment uptake is needed, since eradication of tumor activity cannot be expected after one or two cycles of chemotherapy. Three approaches have been used for quantitative assessment of FDG uptake:

• Kinetic analysis: Patlak analysis, which is easier to perform than classical compartmental analysis and may be used when there's a one-way transport of tracer into tissue, can be used to determine Ki, the influx constant, and to calculate the tumor metabolic rate of FDG.² This may be more sensitive than semiquantitative methods for detecting tumor response, but it also has disadvantages. For one, the method is more onerous and demanding of resources than semiquantitative methods. In addition, when there is marked tumor response to treatment, the Patlak curves are shallow and associated with large uncertainties in Ki. Kinetic constants may also be determined by classical compartmental analysis,³ but this approach is more demanding of resources than Patlak analysis.

• Standardized uptake value (SUV): As the most widely used semiquantitative parameter,⁴ SUV determination involves measuring activity at the target site, with correction for injected dose and body weight. However, this does not correct for inaccuracy of measured dose, which may occur with injected dose extravasation, or for change in tracer excretion or in uptake elsewhere in the body.

• Tumor-to-background ratio (T/B ratio): This ratio has a long history as a semiquantitative parameter, and may have fallen out of favor because of just that--lack of novelty and pizzazz for those who review journal articles and grant applications. This parameter accounts for the same variables as SUV, and also accounts for inaccuracy in dose measurement and change in FDG distribution. It may be due for a comeback.

Patlak analysis is probably too cumbersome for routine clinical application, and semiquantitative parameters are more likely to obtain clinical favor.

Breast cancer
To assess the clinical applications of PET for breast cancer treatment, Wahl et al evaluated 11 patients receiving chemohormonotherapy for primary breast cancer.⁴ PET studies were performed before treatment, and four times during the first three cycles of treatment. FDG metabolism was assessed by means of SUV, and kinetic rate constants also were determined by compartmental analysis.

These researchers demonstrated early decrease in FDG uptake following treatment of responding tumors, as assessed by tumor size, while no such decrease was seen in nonresponding tumors. This work has not been developed further; however, there's probably no potentially curative treatment that can be introduced when nonresponse is demonstrated.

Nonsmall cell lung cancer
For nonsmall cell lung cancer (NSCLC), patients who have evidence of metastasis to mediastinal lymph nodes (N2 disease) may be treated by neoadjuvant chemotherapy and subsequent surgery.⁵ The intention is to render the metastatic tumor nonviable by chemotherapy before surgery. Unfortunately, there has been no way to determine the success of neoadjuvant therapy preoperatively, since change in CT findings is not useful for predicting histologic tumor eradication. For that reason, FDG PET has been evaluated to determine whether glucose metabolic findings can be used to avoid surgery in patients whose tumors have not responded to neoadjuvant therapy, and who are therefore unlikely to benefit from pneumonectomy.

At the Northern California PET Imaging Center in Sacramento, 12 patients with stage IIIA NSCLC underwent PET before and after neoadjuvant chemotherapy, followed by tumor resection. Four patients showed no decrease in lymph node activity with treatment, and all had histologic evidence of viable tumor at surgery. Out of eight patients who showed decrease in lymph node activity at the second PET, five had only necrotic tumor at operation, and three had evidence of continuing viable tumor. These findings suggest that PET may be able to establish lack of response to treatment in a fraction of nonresponding patients, thereby avoiding surgery that would be unlikely to be of benefit.

Evaluating FDG PET for predicting response to chemotherapy has been included in an intergroup trial initiated by the South-West Oncology Group. The principal purpose of the trial is to determine the effect on five-year survival of three cycles of preoperative chemotherapy in patients with stage I NSCLC. FDG PET will be performed before chemotherapy and again after one and three cycles, and change in FDG uptake will be correlated with percent tumor necrosis found at surgery. This study may provide a definitive demonstration of the value or lack of value of FDG imaging in this situation.

Hodgkin's, Non-Hodgkin's lymphoma
For Hodgkin's and Non-Hodgkin's lymphoma, PET has been evaluated for detecting viable tumor in residual post-treatment masses demonstrated by anatomic imaging to evaluate response to a full course of chemotherapy. PET has also been used to evaluate metabolic change after one to two cycles for possible prediction of response to a full course of treatment. Non-Hodgkin's lymphoma is highly responsive to chemotherapy and shows a wide variation in responsiveness to different drug regimens; thus, predicting response to the initial regimen has major management consequences.

De Wit et al⁶ evaluated PET in 28 patients (HD and NHL) with known residual masses by CT, and Jerusalem et al⁷ compared PET and CT in 54 post-treatment patients (HD and NHL). Both groups found that PET findings were highly predictive of subsequent relapse or no relapse.

Early studies by Hoekstra⁸ and by Dimitrialopoulou-Strauss⁹ both demonstrated that patients who failed to respond showed continued high FDG uptake after one to two cycles, while a decrease was seen in responders.

Rohmer et al¹⁰ studied the use of PET before treatment and after one and two cycles of treatment in 11 patients with high-grade NHL for predicting eventual response. They compared SUV and metabolic rate of FDG, measured by the Patlak method, between the pre- and intra-treatment studies. Both SUV and metabolic rate of FDG were significantly lower in the group of patients that remained in complete remission than in the group that relapsed. PET at six weeks (before the third cycle) was found to be a better predictor than PET at one week.

At six weeks, four patients had SUV greater than 2.5, and all relapsed. Seven patients had SUV less than 2.5, and only one relapsed. By CT, eight patients had residual post-treatment masses after two cycles; four showed eventual recurrence and four did not. Of the three patients without a residual mass, one relapsed and two did not. Anatomic imaging had no value in predicting response to therapy.

Head and neck cancer
For head and neck cancer, FDG-PET was evaluated for predicting treatment response by Lowe et al¹¹ in 20 patients undergoing chemotherapy and radiation therapy for stage III/IV head and neck cancer as part of an organ-preservation protocol. When complete remission was not achieved by chemotherapy, surgical salvage was used, so that early detection of nonresponse would have permitted earlier surgery in those patients. Biopsy was used as the criterion standard to evaluate accuracy of PET. PET sensitivity was 90 percent and specificity was 83 percent. Two patients who had positive PET findings and negative initial biopsy proved to be positive on rebiopsy, performed on the basis of the PET result, so that the sensitivity of initial biopsy was actually equal to PET at 90 percent.

Colorectal cancer
Finally, for hepatic metastasis from colorectal cancer, Findlay et al¹² evaluated PET for predicting response to the chemotherapeutic agent 5FU in 18 patients with liver metastasis from colorectal cancer. PET study was performed before treatment and at one to two and four to five weeks after commencement of treatment; results were compared by SUV and T/B ratio.

Thirteen lesions that showed greater than a 15 percent reduction in T/B at four to five weeks all showed subsequent response to treatment by CT or MRI, whereas only one of 10 patients who had reduction of less than 15 percent showed response. By SUV, four out of 14 responding lesions showed an increase at four weeks, and two out of nine nonresponding lesions showed a decrease. These results indicate that T/B ratio may be a more valuable parameter than SUV for serial measurements of tumor FDG metabolic activity.

Overall, FDG PET has been effective in small clinical studies for predicting eventual tumor response to a full treatment regimen in NHL, head and neck cancer and hepatic metastases from colorectal cancer. In all studies, failure to achieve reduction in FDG uptake during treatment has been a strong predictor of nonresponse.

In addition, PET has been shown to be effective in detecting viable tumor in post-treatment residual masses in malignant lymphoma. Early studies also indicate that PET can detect failure of response to neoadjuvant chemotherapy for NSCLC in some patients, thereby avoiding unnecessary tumor resection.

Finally, most evaluations of PET for predicting response to treatment have been based on serial determinations of SUV, but tumor-to-background ratio should be further evaluated for this purpose. Compartmental analysis and Patlak analysis demand too much time and resources for current routine clinical use.

References
1. Cancellos, G.P. (1988). Residual mass in lymphoma may not be residual disease. J Clin Oncol, 6, 931.

2. Patlak, C.S. & Blasberg, R.G. (1985). Graphical evaluation of blood-to-brain transfer constants from multiple time uptake data. J Cereb Blood Flow Metab, 5, 584-590.

3. Phelps, M.E., Huang, S.C., Hoffman, E.L., et al. (1979). Tomographic measurement of local cerebral glucose metabolic rate in humans with [F-18] 2-fluoro-2-deoxy-D-glucose: Validation of method. Ann Neurol, 6, 371.

4. Wahl, R.L., Zasadny, K., Helvie, M., Hutchins, G.D., Weber, B. & Cody, R. (1993). Metabolic monitoring of breast cancer chemohormonotherapy using positron emission tomography: Initial evaluation. J Clin Oncol, 11, 2101-2111.

5. Schilder, R.J., Goldberg, M., Millenson, M.M., Movsas, B., Ragatko, A., Rogers, B. & Langer, C.J. (2000). Phase II trial of induction high-dose chemotherapy followed by surgical resection and radiation therapy for patients with marginally respectable non-small cell lung cancer. Lung Cancer, 27, 37-45.

6. DeWit, M., Bumann, D., Beyer, W., Herbst, K., Clausen, M. & Hossfeld, D.K. (1997). Whole-body positron emission tomography (PET) for diagnosis of residual mass in patients with lymphoma. Ann Oncol, 8 Suppl, 1, 57-60.

7. Jerusalem, G., Beguin, Y., Fassotte, M.F., Najjar, F., Paulus, P., Rigo, P. & Fillet, G. (1999). Whole-body positron emission tomography using F-18-fluorodeoxyglucose for post-treatment evaluation of Hodgkin's disease and Non-Hodgkin's lymphoma has higher diagnostic and prognostic value than classical computed tomography scan imaging. Blood, 94, 429-433.

8. Hoekstra, O.S., van Lingen, A., Ossenkoppele, G.J., et al. (1993). Early treatment response in malignant lymphoma as determined by planar fluorine-18-fluorodeoxyglucose scintigraphy. J Nucl Med, 34, 1706.

9. Dimitrakopoulou-Strauss, A., Strauss, L.G., Goldschmidt, H., et al. (1995). Evaluation of tumor metabolism and multi-drug resistance in patients with treated malignant lymphoma. Euro J Nucl Med, 22, 434.

10. Romer, W., Hanauske, A.R., Ziegler, S., et al. (1998). Positron emission tomography in Non-Hodgkin's lymphoma: Assessment of chemotherapy with fluorodeoxyglucose. Blood, 91, 4464-4471.

11. Lowe, V.J., Dunphy, F.R., Varvares, M., et al. (1997). Evaluation of chemotherapy response in patients with advanced head and neck cancer using [F-18] fluorodeoxyglucose positron emission tomography. Head Neck, 19, 666-674.

12. Findlay, M., Young, H., Cunningham, D., et al. (1996). Noninvasive monitoring of tumor metabolism using fluorodeoxyglucose and positron emission tomography in colorectal cancer liver metastases: correlation with tumor response to fluorouracil. J Clin Oncol, 14, 700-708.

Peter E. Valk, MD, is medical director of Northern California PET Imaging Center in Sacramento, Calif., and professor of molecular and medical pharmacology at UCLA School of Medicine, Los Angeles.