In prior studies, we reported that MRFDG and BF PET measures before and at the midpoint of neoadjuvant chemotherapy predicted response among LABC patients.9
This study expands on our previous studies by exploring the long-term end points of breast cancer recurrence and mortality. We observed that patients whose tumors had increases or small reductions in BF and FDG K1
from pretherapy to midtherapy examinations had elevated recurrence and mortality risks compared with patients with greater reductions in BF and FDG K1
. We also found evidence for higher mortality risk associated with higher BF on midtherapy examinations. Differences in DFS and OS by PET parameters were observed even after adjusting for multiple prognostic factors, such as tumor ER and PR status, size, histology, and pathologic response. Our results suggest that PET data, especially changes in tumor perfusion over the course of neoadjuvant chemotherapy, measured directly by [15
O]water or indirectly by dynamic FDG PET as FDG K1
, provide information distinct from standard markers.
PET as a predictor of patient outcome has been reported for numerous other cancers including sarcoma and head and neck, esophageal, and lung cancers.27-30
For breast cancer in the metastatic setting, qualitatively positive FDG PET tumor uptake after treatment was associated with shorter median DFS or OS.31-33
Prior studies have not examined and compared PET measures of breast tumor BF and FDG tumor kinetics with breast cancer recurrence or mortality risk in the neoadjuvant setting. To our knowledge, our study is the first to show that parameters estimated from kinetic analysis of dynamically acquired PET examinations predict outcome among LABC patients. The standard clinical and pathologic factors also evaluated did not correlate with DFS or OS in this relatively small study, suggesting that quantitative PET imaging provides predictive data independent of these established factors. We observed that persistent MRFDG uptake could indicate tumor resistance to therapy and that greater decreases in BF predicted favorable survival. The standard static measure used for most FDG PET studies, SUV, did not retain predictive value after accounting for other risk factors associated with DFS or OS.
Although a number of tumor and host factors play a role in tumor sustainability, tumor vasculature is necessary for growth and spread. Several different imaging modalities, such as dynamic contrast-enhanced magnetic resonance imaging (MRI), [99m
Tc]sestamibi (MIBI), Doppler ultrasound, and dynamic FDG PET have the ability to assess in vivo tumor BF and vascularity and have shown utility in measuring treatment response.34-40
Our observations of higher relapse and mortality risks associated with higher tumor BF at therapy midpoint parallel our previous work using MIBI imaging41
and MRI findings that evaluated LABC response to antivascular treatment.42
Persistent MIBI uptake, MRI contrast enhancement, and BF in breast tumors after therapy may all indicate the inability of the chemotherapeutic agent to disrupt tumor vasculature, thus allowing continued tumor growth and potential spread, portending a poorer prognosis.
In accordance with previous works that demonstrated a relationship between BF and FDG K1
both before and after therapy,11,12
we observed similar relapse and mortality risks associated with proportionate changes in tumor BF and FDG K1
over the course of chemotherapy. These results suggest that it may be feasible to substitute K1
, the transport rate constant of [18
F]FDG from blood to tissue, for [15
O]water studies, which require an on-site cyclotron.
F]FDG scans acquired in the clinical setting are typically static whole-body images in which semiquantitative tumor uptake measures are dependent on the time interval between tracer injection and scanning,43
which are factors important to replicate when using [18
F]FDG studies for monitoring patient therapy response. Dynamic [18
F]FDG data acquisition is not dependent on image time. Full kinetic analysis provides insights into tumor patterns of glucose metabolism that include transport and phosphorylation measures,44
which are predictive indicators that may not be visualized by static whole-body imaging.
Potential limitations to our study include a relatively small cohort with a recruitment time frame spanning 10 years. Second, although the majority of patients received similar neoadjuvant chemotherapy, there was some heterogeneity of treatment regimens. Third, there was some variability in scan timing and length of treatment before definitive surgery. The difference in treatment lengths and the broad time point range for midtherapy PET examinations is reflective of the ongoing changes in our clinical practice for LABC patients. These findings may not be necessarily generalizable to other populations; further analysis in a larger series is warranted.
Our findings suggest that a small group of breast cancer patients identified by PET experience poor outcome. Early response monitoring would play a critical role for these patients. Prior PET studies indicate that early response monitoring is feasible.7,8
Our study also suggests that targeting tumor vasculature of patients who have resistant tumors may be helpful. Current studies at our institution are evaluating the role of dynamic FDG PET and dynamic contrast-enhanced MRI in early response prediction of antivascular therapies for breast cancer.
Our results suggest that information provided by PET imaging is complementary to standard clinical end points based on surgical pathology.27-30
Therefore, functional imaging may be helpful in clinical trials as an adjunct in measuring tumor response and predicting patient outcome.
Overall, we observed that patients with smaller declines in BF and FDG K1 experienced higher risks of recurrence and mortality that were largely independent of patient and tumor characteristics assessed in this study. Our findings suggest that changes in tumor perfusion over the course of neoadjuvant chemotherapy measured directly by [15O]water or indirectly by dynamic FDG PET are predictive of outcome in LABC patients.