Sequential ^99mTc-Hydrazinonicotinamide-Annexin V Imaging for Predicting Response to Chemotherapy

MATERIALS AND METHODS

RESULTS

Patient data and results are shown in Table 1.

Of 17 patients, 7 responded (3 complete responders and 4 partial responders) to treatment, as assessed with RECIST (Figs. 1 and 2). The baseline ratios of tumor activity to background activity exceeded 1 in 5 responders and were equal to 1 in the other 2 responders. In 6 of the 7 responders, significant changes in the ratios of tumor activity to background activity from the baseline value were evident: a significant increase at 5–7 h and a subsequent decrease at 40–44 h in 4 patients (significantly lower than the baseline value in patient 8), an isolated significant increase at 40–44 h in 1 patient, and a continuing significant increase over time in 1 patient. In patient 6, no changes were seen over time. The lesion under study in this patient had a 1-cm diameter.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1. 

Significant increase in uptake of ^99mTc-HYNIC-annexin V over time in primary breast tumor (arrows) in patient 9.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2. 

Significant increase on scan 2 with subsequent decrease on scan 3 in uptake of ^99mTc-HYNIC-annexin V over time in aortocaval lymph node (arrows) in patient 17 (see ratios of tumor activity to background activity for this patient).

Among the 10 nonresponders, the baseline ratios of tumor activity to background activity were equal to 1 in 8 patients, less than 1 in another patient, and greater than 1 in yet another patient (Fig. 3). In none of the 10 nonresponders did the baseline ratios of tumor activity to background activity change significantly over time.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3. 

Absence of changes in uptake of ^99mTc-HYNIC-annexin V over time in sternoclavicular bone metastasis relative to baseline bone scan (leftmost image) in patient 11.

With only the baseline ratios of tumor activity to background activity (scan 1) and a cutoff value of 1 (>1 for responders and ≤1 for nonresponders), the obtained accuracy for the purpose of separating responders from nonresponders was 82% (14/17 patients). With the proposed sequential imaging protocol (scans 1, 2, and 3) and a 25% change threshold, responders could be separated from nonresponders with a 94% accuracy (16/17 patients). The sensitivity was 86%, the specificity was 100%, the positive predictive value was 100%, and the negative predictive value was 91%.

Of interest was that when the baseline scan (scan 1) was used in combination with either the 5- to 7-h images (scan 2) or the 40- to 44-h images (scan 3) to predict response to therapy, the accuracy was 88% (15/17 patients). With the baseline scan (scan 1) and the 40- to 44-h images (scan 3), however, a significant decrease rather than an increase in the ratio of tumor activity to background activity was evident for patient 1.

DISCUSSION

It is well known from in vitro and preclinical studies that the level of spontaneous cell death in tumors reflects chromosomal instability and that higher chromosomal instability results in higher sensitivity to chemotherapy, which is directly related to chemotherapy-induced cell death (12,13). In agreement with these findings, 8 of the 10 nonresponders in the present study had ratios of tumor activity to background activity at baseline comparable to the background activity, and in none of the patients did ratios of tumor activity to background activity change significantly over time. In contrast, 5 of the 7 responders had ratios of tumor activity to background activity at baseline higher than the background activity, and in 6 responders, a significant increase in the ratios of tumor activity to background activity over time after treatment initiation was evident. The higher-than-background activity at baseline in responders in the present study versus the absence of tumor uptake in responders in the study of Belhocine et al. (9) likely relates to the difference in the kinetics of ^99mTc-BTAP-annexin V (BTAP is [4,5-bis(thioacetamido)pentanoyl]) and ^99mTC-HYNIC-annexin V. Because of the slow blood clearance of ^99mTc-BTAP-annexin V in the study of Belhocine et al., meaningful planar images of low quality could be obtained only at 24 h (4 half-lives) after radiopharmaceutical injection. Unlike ^99mTc-BTAP-annexin V, ^99mTc-HYNIC-annexin V has more favorable clearance characteristics, allowing imaging at 4–6 h after radiopharmaceutical injection (14). By that time, nearly all activity is cleared from the blood and, as shown in a previous report by our group, is unlikely to interfere significantly with measurements of the uptake of ^99mTc-HYNIC-annexin V by tumors (3).

In the 1 responder showing a baseline value of 1 that did not change over time, the lesion was located in the liver and was approximately 1 cm in diameter. In this patient, the results might have been flawed because of a partial-volume effect and spillover from surrounding liver tissue. In 4 responders, the ratios of tumor activity to background activity increased significantly at 5–7 h after treatment initiation and subsequently decreased at 40–44 h. In 1 other responder, the ratio of tumor activity to background activity increased significantly only at 40–44 h after treatment initiation, whereas in yet another responder, this ratio increased significantly at 5–7 h and increased further at 40–44 h. Given the absence of the complete kinetic profile of apoptosis induction in the patients under study, it is impossible to determine whether the increases found were actually peaks of activity or points on a descending or ascending portion of a curve. Also, we cannot exclude the possibility that a wave of apoptosis was missed entirely by the time schedule defined for imaging before the present study was performed. However, on the basis of available preclinical and clinical data outlined here, it is likely that the increase in activity at 5–7 h, evident in 5 of 6 responders, was part of an early peak of cell death or PS expression. The significance of this early peak in ^99mTc-HYNIC-annexin V uptake is currently unclear but may relate to the death of tumor cells at a particularly sensitive phase of the cell cycle, to preapoptotic “stressing” of tumor cells, or to the short-term effects of cytotoxic chemotherapy on vascular endothelial cells.

Evidence that a chemotherapeutic agent directly causes cytotoxicity to the vasculature of a tumor was first reported by Baguley et al. in 1991 (15). In that report, Baguley et al. demonstrated that vinblastine led to greater than 90% cell death of drug-resistant solid tumors within hours but had no effect when the same cells were grown on ascites. Since then, antiendothelial cell effects have been demonstrated in vitro for cyclophosphamide, 5-fluorouracil, and mitomycin C, and short-term antiangiogenic effects have been demonstrated in vitro for vincristine, vinblastine, doxorubicin, mitoxantrone, etoposide, paclitaxel, 6-methylmercaptopurine, tegafur, camptothecin, topotecan, irinotecan, and combretastatin (16,17). The delayed increase in activity evident in 1 patient and the continuing increase in activity compared with the activity on the 5- to 7-h images in yet another patient could represent a second peak of activity assumed to relate to the loss of the bulk of tumor cells through cell death in a fashion similar to that observed in preclinical trials.

It has been suggested that the need for multiple separate injections of ^99mTc-labeled annexin V, as performed in the present study, may be considered a relative drawback and that the use of annexin V derivatives with longer half-lives and adjusted molecular weights may make it possible to capture therapy-induced apoptotic cells over a prolonged period. Along these lines, Wen et al. developed ¹¹¹In-labeled pegylated annexin V (18). In contrast to ¹¹¹In-diethylenetriaminepentaacetic acid (DTPA)-annexin V, ¹¹¹In-DTPA-pegylated annexin V performed much better in terms of visualizing tumor apoptosis. Direct comparative studies of ^99mTc-HYNIC-annexin V and ¹¹¹In-DTPA-annexin V, however, were not performed. As with ^99mTc-BTAP-annexin V, it is possible that the presence of the DTPA linker significantly influences the biodistribution of annexin V, reducing uptake by the tumor relative to that obtained with the HYNIC formulation. In addition, although the increased half-life of ¹¹¹In-labeled pegylated annexin V is likely to increase exposure time and penetration into solid tumors, nonspecific uptake is also likely to increase over time given the lower likelihood of washout. In this regard, the increase in the uptake of ¹¹¹In-DTPA-pegylated ovalbumin, a nonspecific control, by MDA-MB-468 tumors grown in mice at 4 d after paclitaxel treatment was 55.6%; that of ¹¹¹In-DTPA-pegylated annexin V was 96.7% (19). Also, blood-pool activity is likely to remain high, compromising tumor imaging. Further, the use of an isotope with a long half-life means a reduction of the dose injected and accordingly low count statistics. Finally, although it has been suggested that annexin V with long-term circulation may make it possible to capture therapy-induced apoptotic cells over a prolonged period, the occurrence of apoptosis and the subsequent removal of apoptotic bodies by macrophages take only a couple of hours, and it is only an assumption that radiolabeled annexin V stays within macrophages after internalization. Direct comparative studies of ^99mTc-HYNIC-annexin V and ¹¹¹In-DTPA-pegylated annexin V are needed.

CONCLUSION

In the present study, responders to treatment could be separated from nonresponders with a 94% accuracy (16/17 patients) by use of the proposed sequential imaging protocol and a 25% change threshold. However, the number of patients studied was small, and the histologic nature of the tumors explored and the drugs administered were heterogeneous. Prospective studies with the same imaging protocol and more homogeneous groups of patients in terms of tumor histology and drugs administered are needed to confirm these promising results.