Lesion detection by [89Zr]Zr-DFO-girentuximab and [18F]FDG-PET/CT in patients with newly diagnosed metastatic renal cell carcinoma

Purpose The main objective of this preliminary analysis of the IMaging PAtients for Cancer drug selecTion (IMPACT)-renal cell cancer (RCC) study is to evaluate the lesion detection of baseline contrast-enhanced CT, [89Zr]Zr-DFO-girentuximab-PET/CT and [18F]FDG-PET/CT in detecting ccRCC lesions in patients with a good or intermediate prognosis metastatic clear cell renal cell carcinoma (mccRCC) according to the International Metastatic Database Consortium (IMDC) risk model. Methods Between February 2015 and March 2018, 42 newly diagnosed mccRCC patients with good or intermediate prognosis, eligible for watchful waiting, were included. Patients underwent CT, [89Zr]Zr-DFO-girentuximab-PET/CT and [18F]FDG-PET/CT at baseline. Scans were independently reviewed and lesions of ≥10 mm and lymph nodes of ≥15 mm at CT were analyzed. For lesions with [89Zr]Zr-DFO-girentuximab or [18F]FDG-uptake visually exceeding background uptake, maximum standardized uptake values (SUVmax) were measured. Results A total of 449 lesions were detected by ≥1 modality (median per patient: 7; ICR 4.25–12.75) of which 42% were in lung, 22% in lymph nodes and 10% in bone. Combined [89Zr]Zr-DFO-girentuximab-PET/CT and CT detected more lesions than CT alone: 91% (95%CI: 87–94) versus 56% (95%CI: 50–62, p = 0.001), respectively, and more than CT and [18F]FDG-PET/CT combined (84% (95%CI:79–88, p < 0.005). Both PET/CTs detected more bone and soft tissue lesions compared to CT alone. Conclusions The addition of [89Zr]Zr-DFO-girentuximab-PET/CT and [18F]FDG-PET/CT to CT increases lesion detection compared to CT alone in newly diagnosed good and intermediate prognosis mccRCC patients eligible for watchful waiting. Electronic supplementary material The online version of this article (10.1007/s00259-019-04358-9) contains supplementary material, which is available to authorized users.


Introduction
Renal cell carcinoma (RCC) accounts for 2% of all malignancies worldwide, with an estimated 403,262 new cases in 2018. Seventy percent have a clear cell component. Metastatic clear cell (mcc) RCC has a variable course, with a subgroup of patients showing slow disease progression. In those patients, it is safe to observe the course of disease in a period of socalled watchful waiting, avoiding unnecessary side-effects and costs of systemic treatment.
To identify patients eligible for watchful waiting, prognostic schemes such as the International Metastatic Database Consortium (IMDC) risk model have been used to differentiate between patients with a good, intermediate or poor prognosis [1,2]. For staging mRCC, European Society of Medical Oncology (ESMO) guidelines mandate contrast-enhanced computed tomography (CT) of chest, abdomen and pelvis [3].
Previously, an international phase II study in mRCC patients eligible for watchful waiting showed that higher numbers of IMDC adverse risk factors (p = 0.0403) and higher numbers of metastatic disease organ sites (p = 0.0414) were associated with a shorter period of watchful waiting [4]. These results substantiate the clinical value of imaging, which may be further enhanced by molecular imaging with [ 18 F]FDG or emerging radiopharmaceuticals targeting tumor-associated antigens like carbonic anhydrase IX (CAIX) to identify patients in need of urgent systemic or local therapy.

Patients
In this prospective multicenter cohort study, patients aged 18 years and older with histologically or cytologically proven RCC with a clear cell component, recently (<6 months) diagnosed metastases, and a good or intermediate prognosis according to IMDC score [1], were enrolled in the IMPACT-RCC study conducted at four Dutch academic medical centers. A period of watchful waiting for 2 months was considered optional according to treating medical oncologist. Patients who received any previous systemic treatment for RCC in any setting were excluded, but previous radiotherapy and surgery (nephrectomy or metastasectomy) was permitted. Furthermore, patients were excluded in the presence of untreated central nervous system metastases or symptomatic intra-cerebral metastases, pregnant or breast feeding women. Only patients without prior systemic treatment were enrolled, therefore the IMDC criteria 'time from diagnosis to treatment <1 year' was adapted into 'time from primary diagnosis to diagnosis of metastatic disease <1 year'. Watchful waiting was terminated if radiological disease progression was established, in combination with a clinical need to start systemic treatment.

Patient imaging
Patients underwent CT, [ 18 F]FDG and [ 89 Zr]Zr-DFOgirentuximab-PET/CT at the start of the watchful waiting period. Further details on the imaging modalities (acquisition and reconstruction protocols) and the conjugation, radiolabeling and quality control of [ 89 Zr]Zr-DFOgirentuximab are provided in the Supplements.

Image assessment
All CT and [ 18 F]FDG-PET/CT scans were reported according to standard clinical practice by an experienced local radiologist and nuclear physician, respectively. The assessment of CT lesions was performed according to RECIST 1.1 [18]; however, to ensure measurements and documentation of all lesions including non-target lesions of ≥10 mm, CT scans were independently revised by one or two experienced radiologists (E.H.A; T.C.K.). The [ 89 Zr]Zr-DFO-girentuximab PET/CTs were assessed in a central reviewing system to ensure true lesion detection and reproducible inter-observer agreement.
All [ 89 Zr]Zr-DFO-girentuximab PET/CTs were assessed by three expert nuclear physicians independently (W.O.; A.H.B.; O.H.) through online central reviewing system designed by CTMM TRaIT. The three reports were harmonized to one final report by one designated reviewer. In case of different findings, a meeting was organized to reach consensus. The treating physician was blinded for the results of either PET/CT; however, for patient safety reasons, the nuclear physician was allowed to communicate findings that required (local) interventions (e.g. brain metastases).
A tumor lesion was defined visually positive based on anatomical substrate on low-dose CT in combination with [ 18 F]-FDG and/or [ 89 Zr]Zr-DFO-girentuximab-uptake, or solely on prominent, non-physiological antibody-uptake. Quantification of positive lesions as defined by evaluation reports for [ 18 F]FDG and [ 89 Zr]Zr-DFO-girentuximab-PET/CT was performed by drawing regions-of-interest using Inveon Research Workplace software (IRW, version 4.1). The maximum and mean standardized uptake values (SUV) were calculated. SUV max was used for tumor tracer-uptake; SUV mean for measuring uptake in healthy organs and blood pool.

Statistical analysis
To compare the agreement in individual lesion detection between observers, we used dependent pair wise or multiobservers kappa-coefficients with the delta method [19]. Lesion detection rates per imaging modality and combined imaging modalities (CT combined with PET/CT) were estimated and compared (by Wald tests) using mixed effect logistic regression models accounting for within patient and lesionclustering by random intercepts. We evaluated lesion detection rates overall and according to organ sites. Furthermore, we compared the median number of affected organ sites across patients assessed by CT only, or in conjunction with either PET/CT using Wilcoxon signed rank tests.
To assess biodistribution of [ 89 Zr]Zr-DFO-girentuximab, we estimated the average SUV mean per organ and compared variability within and between patients (one-sample T-test). SUV max was evaluated using descriptive methods besides mixed effects linear regression models, taking within patient clustering into account as random intercepts (using intra-class correlation coefficient (ICC) to estimate variation in uptake due to between-patient heterogeneity). These models were also used to assess determinants of tracer-uptake (introduced as fixed effects and compared by Wald tests). SUV max was natural log-transformed to obtain appropriate model fit, resulting in geometric means or percent changes in SUV max as interpretation of fixed effects. We fitted these models under restricted maximum likelihood using Satterthwaite approximations to degrees of freedom. We used the marginal R 2 to estimate the variance in tracer-uptake explained by the fixed effects of these models [20], then fitted under maximum likelihood.

Patients
From February 2015 until March 2018, 42 mccRCC patients were included. All patients had a histopathological diagnosis of the primary tumor, either through (partial) nephrectomy or biopsy in 36 and six patients, respectively. A total of 14 patients had a favourable prognosis. Of the remaining 28 patients, 13 had a predicted intermediate prognosis with one risk factor and 15 patients with two risk factors. This was primarily due to the diagnosis of metastases <1 year after the primary diagnosis (80%) and/or the presence of anaemia (51%). There was no correlation between histology (e.g. mixed vs. pure clear cell) and the estimated prognosis according to IMDC.
All patients without a previous nephrectomy had an estimated intermediate prognosis. In total 57% of all patients presented with metachronous metastases at a median interval of 0.7 (range 0-15) months between primary diagnosis and first metastasis. One patient presented with only sub-centimeter indeterminate lung lesions; therefore, lesions were not included in the analyses. Five others had a negative [ 18 F]FDG-PET/ CT, of whom one plus two other patients had a negative [ 89 Zr]Zr-DFO-girentuximab-PET/CT. In two patients, the [ 18 F]FDG-PET/CT and/or [ 89 Zr]Zr-DFO-girentuximab-PET/ CT revealed brain metastases warranting local treatment with stereotactic radiotherapy and temporary treatment with corticosteroids.
Patient characteristics are shown in Table 1, imaging examples are shown in Fig. 1.

Combination of modalities for lesion detection
With the addition of [ 89 Zr]Zr-DFO-girentuximab-PET/CT and [ 18 F]FDG-PET/CT, lesion detection by CT alone increased from 56% to 91% (95%CI 87-94) and 84% (95%CI 79-88), respectively. Improved lesion detection rate was apparent for all organ sites (Fig. 2) Table 2). Patients were categorized according to the location of their lesions (e.g. lung only; other organ(s) only and both lung and other organs). With the addition of both PET/CTs, two patients were re-categorized from lung only into 'both lung and other organs' based on the additional detected lymph node and bone lesions (Table 1).

Quantitative analysis of [ 89 Zr]Zr-DFO-girentuximab and [ 18 F]FDG-uptake
In normal tissues the highest [ 89 Zr]Zr-DFO-girentuximab-uptake was observed in healthy liver (geometric mean SUV mean 6.7 (95%CI 6.4-7.3), lowest in healthy lung (geometric mean SUV mean 1.      A patient's prognosis is estimated based on the number of involved organs on CT, total disease burden and period of watchful waiting, rather than the number of lesions [4,21]. In our study population 33% of the patients present with a predicted good prognosis mRCC and 43% of patients with synchronous metastases. This is comparable to previous datasets and reflects daily clinical practice [4]. Patients with lung-only metastases are thought to have a better prognosis than other involved organ sites such as liver and bone [22,23]. In our study population, based on CT only, seven patients (17%) presented with lungonly metastases. This number was revised after the addition of PET/CT because of the detection of additional bone and lymph node lesions by PET/CT in two patients. Furthermore, two patients were diagnosed with brain metastases by [ 89 Zr]Zr-DFOgirentuximab PET/CT that required local treatment.

Determinants of tracer-uptake
Overall, the median number of two involved organs per patient as determined by CT alone increased to three per patient with the addition of PET/CT (range 1-7; p < 0.005), even without adjusting for the limited CT field-of-view. This is largely attributed to the detection of more soft-tissue and bone lesions, a well-known limitation of CT due to less soft tissue contrast and the limited ability to detect (non-lytic) bone lesions. This limited increase in the number of involved organ sites with the addition of PET/CT questions its additional value, since solely an increase in detection lesions will not lead to the implementation of [ 89 Zr]Zr-DFO-girentuximab or [ 18 F]FDG-PET/CT to our standard work-up. However, The interpretation of involved organ sites in all three modalities was challenging, especially considering the limitation of each modality. For example, spatial resolution is lower with PET/CT compared to CT, resulting in a partial volume effect affecting small (<2 cm), low-contrast lesions both visually and quantitatively [24]. CT can detect sub-centimeter or indeterminate pulmonary nodules and lymph nodes, although distinguishing nonspecific from small metastatic lesions with CT is notoriously difficult. Based on studies of pulmonary metastases in RCC and RECIST 1.1 criteria, we used a diameter cut-off of 10 mm and in lymph nodes 15 mm to prevent overestimating of the number of detected lesions [25,26]. This ultimately reduced the number of (small) lung and lymph node lesions detected by CT, thereby underestimating the overall lesion detection by CT.
All lesions visible on either CT, [ 89 Zr]Zr-DFO-girentuximab or [ 18 F]FDG-PET/CT were defined as metastases, which introduces potential bias and a risk of possible false-positive. Despite the high specificity of [ 89 Zr]Zr-DFO-girentuximab to visualize primary and metastatic ccRCC-lesions expressing CAIX [5,[13][14][15]27] ]Zr-DFO-girentuximabuptake strongly depend on the organ where the lesion is localized, which is also previously described for [ 89 Zr]Zrbevacizumab uptake in mccRCC [28]. Overall, highest [ 18 F]FDG and [ 89 Zr]Zr-DFO-girentuximab SUV max values were visualized in metastases of the adrenal gland and the kidney. In the six patients without previous nephrectomy, the highest SUV max value was measured in the metastatic lesions and not in the primary tumor. Organ-specific characteristics influence [ 89 Zr]Zr-DFO-girentuximab-uptake, e.g. presence of stromal and immune cells, stroma and/or vasculature affecting perfusion. This is illustrated by the notably high SUV max values of [89Zr]Zr-DFOgirentuximab uptake in adrenal gland lesions as compared to other lesion sites, e.g. lung (median SUV max 69.9 and 9.4, respectively).
Depending  [27]. Furthermore, the quantification of tracer-uptake in both PET-imaging modalities offers a better understanding of the heterogenic study population [4]. Combining anatomical imaging techniques with functional imaging techniques targeting glucose metabolism and CAIX expression offers a better representation of the heterogeneity by visualizing whole body tumor nature and active metabolic processes (e.g. glycolysis, GLUT-1-expression) [8].
Upon completion of the follow-up data of all patients included in the IMPACT-RCC study, we will analyze whether [ 89 Zr]Zr-DFO-girentuximab and [ 18 F]FDG-PET/CT contributes to a better prediction of the course of disease during watchful waiting in good and intermediate prognosis mccRCC patients.

Conclusions
The addition of [ 89 Zr]Zr-DFO-girentuximab-PET/CT and [ 18 F]FDG-PET/CT to CT increases lesion detection compared to CT alone in newly diagnosed good and intermediate prognosis mccRCC patients eligible for watchful waiting. The quantitative analyses of 89 Zr-DFO-girentuximab and [ 18 F]FDG-uptake can be relevant in clinical practice, as sitespecific heterogeneity may require a different treatment approach.
Financial support This work was supported by the Dutch Cancer Society (Alpe d'HuZes Grant RUG 2012-5400).