Total Tumor Uptake and Absorbed Dose of 177Lu-Lilotomab Satetraxetan in a First in Human Trial for Relapsed Non-Hodgkin Lymphoma - Are We Hitting the Target?

Purpose: 177 Lu-lilotomab satetraxetan, a novel CD37 directed radioimmunotherapy (RIT), has been investigated in a rst-in-human phase 1/2a study for relapsed non-Hodgkin lymphoma (NHL). Absorbed dose for all tumor tissue in the body is a crucial parameter in RIT which has traditionally been challenging to calculate. The aim of this study was to investigate the correlations between baseline FDG PET/CT and posttreatment SPECT/CT parameters, absorbed dose-response relationships and clinical responses. Materials and methods: A total of 15 patients with different pre-treatment and pre-dosing regimens were included. 177 Lu-lilotomab satetraxetan was administered at dosage levels of 10, 15 or 20 MBq/kg. Total radioimmunoconjugate tumor volume (tRTV), total radioimmunoconjugate lesion uptake (tRLU) and total tumor absorbed dose (tTAD) were calculated from posttreatment SPECT/CT. The measured uptake values and absorbed doses were normalized for dosage when appropriate. For some of the analyses, the cohort was divided into low (arm 1) and high (arm 4+5) non-radioactive lilotomab pre-dosing groups. tMTV and tTLG were calculated from FDG PET/CT performed at baseline, 3 and 6 months after RIT, and the percent change for these parameters calculated ( ∆ tMTV 3months , ∆ tTLG 3months and ∆ tMTV 6months , ∆ tTLG 6months ). Clinical responses were evaluated at 6 months. Results: tTMV and tRTV were signicantly correlated (p<0.01). A correlation was also found between tTLG and tRLU (p<0.01). Correlations were not observed between baseline tTMV and tTAD. Decreases in ∆ tMTV and ∆ tTLG were signicantly higher at PET 3months for patients receiving tTAD ≥ 200cGy compared to patients receiving lower tumor absorbed doses (p=.03 for both). Also, signicant decreases in ∆ tMTV 3months , ∆ tTLG 3months and ∆ tMTV 6months , ∆ tTLG 6months were observed with increasing tTAD in the high lilotomab patient group. Similarly, responders (patients with complete remission and partial remission) had higher mean tTAD compared to non-responders (stable disease and progressive disease). This was statistically signicant in the high lilotomab group. Across the entire population, all non-responders had tTAD < 200cGy, and all patients with tTAD ≥ 200cGy were responders. Conclusion: This work indicates that 177 Lu-lilotomab satetraxetan targets FDG avid lesions, and that increasing baseline (tMTV) does not have a decreasing effect on the total tumor absorbed dose (tTAD). The patient group receiving a higher amount of lilotomab pre-dosing demonstrated an absorbed dose–response relationship. Similar results were not observed in the low lilotomab group, which were expected since an overall very good response rate could mask such a relationship for this group. Regardless of pre-dosing, a mean absorbed dose to the total tumor tissue (tTAD) limit of 200cGy may prove valuable to separate clinical non-responders from responders.


Introduction
Individualized treatments in modern oncology demand accurate measurement of the amount of pharmaceutical reaching the target. Pharmacokinetic (PK) studies are often applied as indirect methods to theoretically determine the distribution of the pharmaceutical both in normal tissue and tumor. However, radiolabeled targeted therapies enable the direct measure of the amount of radiopharmaceutical accumulating in normal tissue and tumor. Such methods have become more precise with advances in imaging technologies and software solutions to calculate the information made available by imaging. Furthermore, accurate measuring of tumor burden before treatment has also become more feasible with advances in imaging technologies and is proposed as part of individualized therapy strategies. In recent years, tumor volume measurement has gained increased interest as a parameter to guide individual dosage adjustments [1].
Targeted therapies like monoclonal antibodies (mAbs) administered as single agents or in combination with other treatments have changed the course of non-Hodgkin lymphoma (NHL). Clusters of differentiation (CD) 20 targeting mAb, rituximab, was the rst of its kind to be approved and is still the most widely used. Variations in response were reported when rituximab was given as single agent since its introduction [2,3]. Several studies in early 2000´s investigated if this variation may be explained by factors like tumor burden, antigen concentration in tumor, circulating antigens and genetic factors [4][5][6]. At rst, tumor burden was measured by computer tomography (CT) and utilized methods like sum of perpendiculars of all lesions, sum of perpendiculars of target lesions or longest diameter of the largest involved node. With the introduction of metabolic tumor volume (MTV) and total metabolic tumor volume (tMTV) as 18 F-FDG PET/CT (FDG PET hereafter) parameters [7], measuring tumor volumes has become easier and more precise. Another FDG PET parameter, total lesion glycolysis (tTLG), helps characterize tumor biology as glucose consumption, but it is still a subject to debate on how it may be applied in treatment planning or evaluation.
Radioimmunotherapy (RIT) works both as targeted radiotherapy and immunotherapy, and this makes it possible to establish image proof of radioimmunoconjugates successfully targeting the viable tumor mass by measurements of the amount of uptake, volume of uptake, and tumor absorbed dose. Alongside with development of rituximab, mAbs with ß-emitting radionuclides have also been tested in clinical trials [8][9][10][11], Methods have been proposed to measure the patient mean tumor absorbed dose for 131 I-tositumomab [12][13][14].
However, to our knowledge, no studies have been conducted with RIT against lymphoma to investigate the impact of baseline tMTV/tTLG on total radioimmunoconjugate uptake in all tumor tissue and the patient mean tumor absorbed doses for the total tumor volume (from here on referred to as total tumor absorbed dose -tTAD). Furthermore, no previous studies have included absorbed dose versus treatment induced changes in tMTV / tTLG for this patient group. 177 Lu-lilotomab satetraxetan or Betalutin ® (Nordic Nanovector ASA, Oslo, Norway) has been investigated in the rst-in-human phase 1/2a study LYMRIT-37-01 for treatment of relapsed CD37 + indolent NHL [15]. Different combinations of non-radioactive pre-dosing and pre-treatments were explored in ve different arms of the phase 1 dose escalation part of this study. In the current sub-study of LYMRIT-37-01 we aimed to investigate 177 Lulilotomab satetraxetan radioimmunoconjugate uptake parameters on the whole body level, and evaluate the impact of baseline tMTV / tTLG on total radioimmunoconjugate lesion uptake (tRLU) and total tumor absorbed dose (tTAD). Furthermore, the potential therapeutic effect of tTAD, measured as change in FDG PET parameters (ΔtMTV 3months , ΔMTV 6months and ΔtTLG 3months , ΔtTLG 6months ) and clinical response after 6 months, were then analyzed.

Patient Characteristics and Treatment
Fifteen patients with relapsed B-cell indolent NHL from the multicenter phase 1/2a LYMRIT-37-01 trial led by Oslo University Hospital were included in this work. Table 1 shows patient characteristics. Only patients from our center, eligible for dosimetry, were included to assure image standardization. CD37 status of patients were con rmed by immunohistochemistry. Histological subtypes were follicular lymphoma grade I-IIIA and mantle cell lymphoma. The LYMRIT-37-01 trial was approved by the regional ethics committee, and all patients had signed an informed consent form.  1). Patients were also grouped further based on pre-dosing, de ning arm 1 with 40 mg lilotomab as the "low lilotomab group" and arms 4 and 5 receiving 100mg/m2 or 60mg/m2 as the "high lilotomab group", respectively ( Fig. 1). We assumed that the modest difference between arm 4 and 5 regarding lilotomab pre-dosing would not signi cantly in uence post-therapy SPECT and follow-up PET parameters. PET 3months and PET 6months if uptakes were higher than liver uptake de ned by PERCIST criteria [17]. Otherwise, they were registered as zero. Figure 2a illustrates the entire tumor uptake volume under the diaphragm at PET baseline in one of the patients. Changes in these parameters from baseline to PET 3months and PET 6months were calculated as percent reduction from baseline value, de ned as ΔtMTV 3months , ΔMTV 6months and ΔtTLG 3months , ΔtTLG 6months . Negative values represent increase in tMTV or tTLG. All measurements were performed by an experienced nuclear medicine physician. Two patients did not undergo PET 3months and PET 6months (one of these patients did not undergo contrast enhanced CT (CeCT) either) and one patient did not undergo PET 6months because of progression.

SPECT/CT imaging and quanti cation
Patients underwent SPECT/CT at day 4 and day 7 post injection (p.i.) of 177 Lu-lilotomab satetraxetan in arm 1, and at day 1, 4 and 7 p.i. in arm 4 and arm 5 ( Fig. 1). SPECT/CT scans were acquired with a dual-head Symbia T16 (Siemens Healthineers) camera. Scanner protocol and reconstruction parameters have been described previously [18]. SPECT/CT data were segmented using the software program PMOD (version 3.6; PMOD Industries) and later post-processed with in-house written python software (version 2.7). tRTVs representing tumor volumes with 177 Lu-lilotomab satetraxetan uptake were determined on the day 4 and 7 SPECT/CT scans by a semi-automatic approach. An initial manual segmentation was performed by a nuclear medicine specialist to exclude physiological uptake in normal tissue in close proximity to lesions. Then, a thresholding with a 26 % cut-off based on the voxel with the highest uptake in the initial segmentation was carried out. This threshold was chosen after a visual optimization. The segmentation was done individually for the day 4 and day 7 scan. The nal segmentations were visually veri ed by side-by-side comparison with the FDG PET (Fig 2a-b). The tRLU was de ned as the total activity inside the tRTV. Cumulative activity concentration was calculated by assuming a mono-exponential wash-out of the activity in the tumors and analytically integrates this curve. tRLU Day0 was also calculated from this curve. Total tumor absorbed dose, de ned as tTAD was calculated from the tRLU curve and the tRTV, by assuming a local dose deposition of all electron radiation particles, equating to 0.0853 Gy/(MBqhrs/ml) [19].

Response assessment
Responses were assessed by FDG PET and CeCT at 3 and 6 months after treatment according to the Cheson criteria [20,21] de ned as complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Bone marrow biopsy was performed to con rm CR if a bone marrow biopsy at baseline was positive. PD was con rmed by CT only.

Statistics
Spearman-rank correlation tests were performed to investigate relationships between PET and SPECT parameters and between SPECT parameters and changes in PET parameters. A signi cance level of 5% was used. Statistical analyses were carried out only for groups of more than two patients. The Mann-Whitney-U test was performed to test differences between groups. A null-hypothesis of equal populations with a rejection level of 5% was set. The boxplots show median values, interquartile ranges, and points lower or higher than 1.  Table 3 and 4, and individual values are provided in supplementary table 1.   (Fig. 3c). However, radioimmunconjugate activity concentration (expressed as tRLU dosage /volume) did not correlate signi cantly with SUV mean (r = .48 p = .07), indicating that consumption of glucose and CD37 expression on tumor cells does not correspond (Fig. 3d).
Three different dosage levels of radioimmunoconjugate were tested in this study. We observed that the total tumor absorbed dose (tTAD) increased with increasing 177 Lu-lilotomab satetraxetan dosage levels. The differences in tTAD between patients treated with 15 MBq/kg and 20 MBq/kg were not signi cant (p = .37) (Fig. 4a). The 10 MBq/kg group was not included in this analysis because it only contained two patients. Signi cant correlations between tTAD and tRLU ( Fig. 4b) and tRLU day7 , were found (both p < .01). This is expected as tTAD was calculated based on the uptake measured at day 4 and day 7 SPECT.
The mean percent of injected activity accumulated in tumor volume calculated from the 177 Lu-lilotomab satetraxetan uptake at day 0 was 3.0% (range 0.18%-9.63%). Increasing percent of injected activity in tumor with increasing tumor volumes (tMTV baseline ) was observed but this was not signi cant (r = .48 p = .07).
Patient weight and BSA are parameters which may affect biodistribution of pharmaceuticals as proposed by PK studies [22][23][24][25]. Our analysis did not show any correlations between neither patient weight nor BSA versus tTAD dosage (p = .34 and p = .50 respectively), and activity concentration in tumor (tRLU dosage /volume) (p = .59 and p = .66, respectively).
Total tumor absorbed doses normalized by 177 Lu-lilotomab satetraxetan dosage levels (tTAD dosage ) did not differ signi cantly across low and high lilotomab groups (p = .61) but was slightly higher in the high lilotomab group.
Changes in metabolic tumor volumes and glucose consumption at PET 3months , ΔtMTV 3months and ΔtTLG 3months , were signi cantly higher for the tTAD ≥ 200cGy group compared to the patient group receiving less than 200cGy (p = .03 for both) (Fig. 6a-b). A similar correlation was shown at PET 6months , ΔtMTV 6months and ΔtTLG 6months , but did not reach signi cance (p = .07 for both) (Fig. 6c-d).
Percent changes in ΔtMTV 3months , ΔtMTV 6months , and ΔtTLG 3months and ΔtTLG 6months , were statistically signi cantly correlated with increasing tTAD in the high lilotomab group (r = .85 p < .01, r = .82 p = .02, r = .87 p < .01 and r = .86 p = .01, respectively), but not in the low lilotomab group (r = .23 p = .70, r = .55 p = .33, r = .55 p = .33 and r = .55 p = .33 respectively) (Fig. 7a-d; decrease illustrated as positive values and increase illustrated as negative values). The low lilotomab group had an overall better response that may contribute to this lack of correlation.
Five patients had CR, two had PR, ve had SD and two had PD ( Fig. 8a and supplementary table 1). tTAD was statistically signi cantly higher in responders (CR + PR) compared to non-responders (SD + PD) in the high lilotomab group (p = .04). This analysis was not carried out in in the low lilotomab group because only two patients in this group were non-responders (Fig. 8b). Large variations in tTAD were observed in responders in low lilotomab group (range 40-420 cGy) (Fig. 8b) (Supplementary table 1). Across the entire cohort, independent of amount of pre-dosing, all non-responders had tTAD < 200cGy (Fig. 8a and c) and all tTAD ≥ 200cGy were responders (Fig. 8a-c).

Discussion
In this era of precision medicine and personalized therapy it is imperative to explore the best way of delivering a treatment with precise dosing tailored for each individual patient. Although time-consuming, tumor and normal tissue dosimetry is a crucial part of targeted radiotherapies, and should be standard both in the clinical setting and in trials according to Council Directive 2013/59/EURATOM [26]. Radioimmunoconjugate uptake determined by post-therapy SPECT derived metrics is an accurate method of analyzing the amount of radioactivity accumulating in tumor; an option unavailable for non-radioactive mAb treatments. Hence, in this sub-study of LYMRIT-37-01, the total amount of 177 Lu-lilotomab satetraxetan accumulated in tumor (tRLU), total tumor uptake volume (tRTV) and total tumor absorbed doses (tTAD) were calculated from post-therapy SPECT/CT. We found that tRTV and tRLU dosage correlated signi cantly with tMTV and tTLG respectively, indicating that 177 Lu-lilotomab satetraxetan targets FDG avid tumor tissue without a reduction in tumor uptake in larger tumor volumes. Furthermore, especially for the high lilotomab group, tTAD showed an impact on both ΔtMTV and ΔtTLG, and on clinical response.
We interpret the strong correlation between baseline tMTV and both tRTV (Fig. 3a) and tRTV day7 (data not shown) as a validation of our method of measuring tRTV. This correlation may be expected as we customized the SPECT threshold side-by-side with FDG PET images to determine a value for calculation of a nal radioimmunoconjugate tumor volume. Still, anatomical agreement of uptake regions is required for such an approach to yield satisfactory results. While the xed threshold of 26% of the maximum uptake for calculating tumor volumes on SPECT doesn't provide a regression slope of exactly one versus tMTV, it provided the best visual agreement. Future studies are needed to investigate whether this threshold can be applied to other targeted radiotherapies.
Despite the strong correlation between tTLG and both tRLU dosage (Fig. 3b) and tRLU dosage day7 , no correlation between activity concentration de ned by tRLU dosage /volume and SUVmean (calculated across the total tumor tissue) was found (r = .48 p = .07) (Fig. 3d). Thus, the tTLG vs tRLU dosage correlation can possibly be attributed to the fact that these parameters were derived from their respective volumes rather than a similarity between consumption of glucose and CD37 expression on these cells. However, this still supports that 177 Lu-lilotomab satetraxetan successfully targets the viable tumor cells in the volume of interest determined from baseline FDG PET.
Standard PK methods without molecular imaging based support assessed to theoretically calculate the amount of a pharmaceutical reaching the tumor volumes is not straight forward, mainly because of changes in biodistribution outside blood compartment as shown by Stokke et al. [27]. Direct image-based measurement of the amounts accumulating in the tumor mass would be preferable for all treatments. However, while this is feasible for targeted radiotherapies where it also enables the calculation of tTAD, it is still a grossly underutilized method. From such measurements, several interesting ndings were derived for 177 Lu-lilotomab satetraxetan in this work. A strong correlation between tRLU dosage and tRTV (r = .75, p < .01) implicates that increasing tumor volumes do not reduce 177 Lu-lilotomab satetraxetan accumulation in tumor (Fig. 3c). This was also demonstrated by increase in mean percentage of injected activity reaching the tumor volumes with increasing tMTV, although this was not signi cant (r = .48, p = .07). In addition, lack of correlation between tMTV and tTAD dosage (Fig. 5a, [27]. Thus, the approach using WB absorbed doses is probably not precise enough to re ect the amount reaching the tumor and organs at risk for 177 Lu-lilotomab satetraxetan. Application of tTLG baseline in treatment planning or changes in this parameter to evaluate response during and after treatment in lymphoma has been proven useful [28,29]. In our study, lack of correlation between baseline tTLG and tTAD dosage indicates that absorbed dose cannot be predicted by FDG uptake intensity at baseline FDG PET (Fig. 5b). 177 Lu-lilotomab satetraxetan activity concentration in tumor (tRLU dosage /volume) did not correlate with SUV mean neither, as discussed above, and in support of the assumption that FDG uptake intensity does not necessarily correlate with CD37 expression in tumor. Further studies are needed to assess the role of heterogeneity in tumors in regard to both FDG and radioimmunoconjugate uptake and how they overlap.
We have previously investigated lesion-based tumor absorbed doses and dose-response relationships, with analyses including 1-5 lesions per patient [30]. The criteria for lesion inclusion were then strictly de ned to allow for precise individual dosimetry of each tumor. Signi cant intra-patient variations were observed and absorbed dose-response relationship at lesion level could not be demonstrated based on changes in FDG PET parameters and Deauville 5-point-scale [30]. By measuring tTAD we here averaged out intra-patient variations and most importantly avoided possible selection bias. While it can be argued that mean absorbed dose is not an adequate metric, and that local low-dose areas are relevant for the overall response, this parameter has been demonstrated as a signi cant predictor for 131 I-tositumomab treatment [13,12]. Mean tTAD in our study was 170 cGy (median 130cGy). This is lower than the median value of between 341 and 275 cGy reported with 131 I-tositumomab (Bexxar) by Dewaraja et al. [13,12]. Methodologies applied in these two studies are partly comparable to ours, although the CT-driven approach for tumor delineation, performed for 131 I-tositumomab, can potentially result in a lower mean tumor absorbed dose (i.e. tTAD) compared to our current method which may exclude tumor tissue with very low uptake.
Based on the proposal by Dewaraja et al [13], we decided to pursue a 200cGy tTAD threshold by investigating the changes in FDG PET parameters and response status strati ed by this limit in our population. ∆tMTV 3months , ∆tTLG 3months , ∆tMTV 6months and ∆tTLG 6months were higher in tTAD ≥ 200cGy group and this difference was signi cant for ∆tMTV 3months and ∆tTLG 3months (Fig. 6a-b), indicating that there is indeed an absorbed dose response correlation also for 177 Lu-lilotomab satetraxetan and that the same threshold can be applied. All four patients with tTAD ≥ 200cGy had ∆MTV 3months ≥ 90%. Variations in response in the lower tTAD (< 200cGy) group was larger. While the patient with lowest tTAD (37cGy) had ∆MTV 3months = 96% and ∆MTV 6months = 89%, a patient with progression (∆MTV 6months = -77%) had tTAD = 100cGy. One of the patients with progressive disease was the only mantle cell lymphoma in our study with tTAD = 77 cGy. Even though mantle cell lymphomas have been characterized as radiosensitive [31], like follicular lymphomas, this patient unfortunately did not respond to 177 Lu-lilotomab satetraxetan treatment. There are few patients in our study and these dissident ndings may be random, but it is likely that absorbed doses ≥ 200cGy gives a more predictable effect, whereas the response to lower absorbed doses (< 200cGy) may be more dependent on individual radiosensitivity.
When analyzing the effect of pre-dosing on absorbed doses we observed a slight but not signi cantly higher tTAD dosage in high lilotomab group. Interestingly, mean ΔtMTV 3months , ΔMTV 6months , ΔtTLG 3months and ΔtTLG 6months were lower in this group despite slightly higher tTAD (Table 3). A clear dose-response relationship was illustrated for this group, with higher tTAD inducing statistically signi cant metabolic tumor volume shrinkage and reduction in lesion glycolysis (Fig. 7a-d). On the contrary, the low lilotomab group with slightly lower tTAD dosage , had higher mean ΔtMTV 3months , ΔMTV 6months , ΔtTLG 3months and ΔtTLG 6months (Table 3). Doseresponse relationships could not be demonstrated in this group (Fig. 7a-d). This is expected since the overall very good response rate could mask a possible dose-response relationship. Why such a difference in response as higher mean ∆tMTV 3months , ∆tTLG 3months , ∆tMTV 6months and ∆tTLG 6months was observed in low lilotomab group and whether other factors that may in uence the response are still open questions.
The LYMRIT 37 − 01 PK study demonstrated an increase in blood activity adjusted exposure (area under the curve) with higher lilotomab pre-dosing levels. According to this analysis, arm 4 (high lilotomab) demonstrated highest exposure, lowest clearance and longest biological half-life of 177 Lu-lilotomab satetraxetan, slightly higher than arm 1 (low lilotomab) [15]. Furthermore, lower bone marrow and spleen absorbed doses in arm 4 [27] in addition to higher blood exposure shown by PK [15] indicates that more 177 Lu-lilotomab satetraxetan is available for tumor uptake. This proposed effect was supported in this study by slightly higher tTAD dosage in the high lilotomab group, even though this was not signi cant. Larger tTAD dosage variations were also observed in the high lilotomab group, in line with our previous lesion based tumor absorbed dose analysis [30].
Evaluation of response versus tTAD also supports the assumption of absorbed dose-response relationships and a 200 cGy threshold. Patients with CR had large variations in tTAD (range 69.5-418.3 cGy) (supplementary table  1), while all patients with SD or PD had tTAD < 200cGy (Fig. 8a and c). Only two patients had PR; one just above a tTAD of 200 cGy and one below. Notably, all patients with tTAD ≥ 200 were responders, whereas all nonresponders had tTAD < 200cGy (Fig. 8c). Based on this analysis, we propose that above a threshold of 200cGy CR is probable, while for < 200cGy large variations in response should be expected. Our methodology for tTAD can exclude tumor volumes with low uptake (as discussed above), however, it ensures that we never overestimate the patients' mean tumor absorbed doses. This means that our conclusions with respect to the 200 cGy limit are conservative and can be safely employed regardless of methodology. If we were to apply a different approach, resulting in lower tTADs, this would not misplace any < 200 cGy patients in the ≥ 200 cGy group (only CR). Hence, the observation that all non-responders had tTAD < 200cGy would also hold true using a different approach. When comparing responders and non-responders in low-and high lilotomab groups, a similar pattern as for the PET response evaluation was revealed. tTAD was statistically signi cantly higher in responders (CR + PR) compared to non-responders (SD + PD) in the high lilotomab group (p = .04). In the low lilotomab group the response rates were higher, and there were only two patients with SD + PD (Fig. 8b). The reason for the difference between the high and low lilotomab groups is not clear, as discussed above, but regardless of pre-dosing all nonresponders had tTAD < 200cGy.
We observed increasing tTAD with increasing 177 Lu-lilotomab satetraxetan dosage levels in this study (Fig. 4a), however, the differences were not signi cant between 15 MBq/kg and 20 MBq/kg groups (p = .37) (the 10MBq/kg group was not included in this analysis because the group consisted of only two patients). This illustrates that increasing the amount of activity administrated will not necessarily increase the absorbed dose signi cantly as this value will also depend on patient-speci c uptake and kinetics. While ΔtMTV 3months , ΔMTV 6months , ΔtTLG 3months and ΔtTLG 6months did not vary between the two dosage levels (p = 1, p = .71, p = 1 and p = .71 respectively), there was a difference for these parameters according to tTAD (threshold 200cGy, as discussed above, p = .03 for both ΔtMTV 3months and ΔtTLG 3months , and p = .07 for both ΔMTV 6months and ΔtTLG 6months ) (Fig. 6a-d). This nding indicates that response does not necessarily directly rely on dosage levels, and that absorbed dose can be further investigated as a solitary predictor.

Conclusion
In this study 177 Lu-lilotomab satetraxetan total tumor absorbed doses were calculated and a rarely seen absorbed-dose-response relationship was revealed in the high lilotomab pre-dosing group. While similar results were not observed for the patients receiving lower amounts of lilotomab, this was probably since this group had an overall very good response rate. Increasing tumor burden did not have a decreasing effect of on availability of mAbs, indicating that the amount of 177 Lu-lilotomab satetraxetan given was su cient for all tumor volumes studied. However, further studies are needed to establish this in a patient population with a larger range of volumes. Prediction of CR with tumor absorbed doses ≥ 200cGy is reasonable, while large variations of response should be expected with tumor absorbed doses < 200cGy.
We argue that well-designed dosimetric studies are largely underutilized as the most direct method to measure the uptake of targeted therapies. This provides valuable information to determine the optimal radioimmunoconjugate dosage levels and pre-dosing regimens to attain the highest possible absorbed dose to tumor while maintaining acceptable absorbed doses to normal tissues.

Declarations
Acknowledgments We thank the personnel at the Nuclear Medicine department at Oslo University Hospital for technical assistance with the acquisitions. Stine Nygaard, study nurse at the Department of Oncology, is also greatly acknowledged.
Authors' contributions All authors contributed to design and draft of the manuscript. All authors read and approved the nal manuscript.
Compliance with ethical standards Con ict of interest Arne Kolstad were both in part supported by grants from the Norwegian Cancer Society. Arne Kolstad is member of the Scienti c Advisory Board of Nordic Nanovector ASA. Jostein Dahle is an employee and shareholder of Nordic Nanovector ASA. Ayca Løndalen has no con ict of interest. Johan Blakkisrud has no con ict of interest. Mona-Elisabeth Revheim has no con ict of interest. Caroline Stokke has no con ict of interest.
Ethical approval and informed consent All procedures performed were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study. The zero-hour time point on the grey time line indicates administration of 177Lu-lilotomab satetraxetan. The current study included arms with three different pre-dosing regimens given 1-3 hours before 177Lu-lilotomab satetraxetan injection. Based on pre-dosing, patients were here divided into two groups as indicated; low and high lilotomab. Pre-treatment regimens were given 28 and 21 days before or 14 days before the radioimmunoconjugate. FDG PET was performed as baseline investigation and at 3 and 6 months Illustration of FDG PET/CT and 177Lu-lilotomab satetraxetan SPECT/CT images and uptake agreement for tumors. a PETbaseline with all tumor volumes with uptake higher than liver uptake segmented. b All tumor volumes at day 4 SPECT after a 26% of the maximum uptake threshold was applied. Physiological uptake was removed from both PET and SPECT Figure 3 a 177Lu-lilotomab satetraxetan uptake volume tRTV plotted against tMTVbaseline. A statistically signi cant correlation between the volumes tMTVbaseline and tRTV supports the validity of our method to measure tRTV. b a Increasing absorbed dose to the total tumor volume, tTAD, was observed with increasing 177Lu-lilotomab satetraxetan dosage levels. However, the differences in tTAD were not signi cant between the groups (15 MBq/kg vs 20 MBq/kg; p=.37, 10 MBq/kg group excluded because of few patients). b tTAD plotted against tRLU. As expected, there was a signi cant correlation between these parameters Fig. 5 a tTADdosage plotted against tMTVbaseline. There was no signi cant correlation between baseline tMTV and tTADdosage (r=.30, p=.28), implicating that increasing tMTV did not have a reducing effect on tTAD. b tTADdosage plotted against tTLG. tTLG did not correlate with tTADdosage (r=.42, p=.12). This indicates that absorbed dose cannot be predicted by the FDG uptake at baseline FDG PET Figure 5 a tTADdosage plotted against tMTVbaseline. There was no signi cant correlation between baseline tMTV and tTADdosage (r=.30, p=.28), implicating that increasing tMTV did not have a reducing effect on tTAD. b tTADdosage plotted against tTLG. tTLG did not correlate with tTADdosage (r=.42, p=.12). This indicates that absorbed dose cannot be predicted by the FDG uptake at baseline FDG PET Figure 6 a ∆tMTV3months, b ∆tTLG3months, c ∆tMTV6months and d ∆tTLG6months were higher for patients with absorbed doses of over 200 cGy to the total tumor volume (tTAD). This difference was signi cant for ∆tMTV3months and ∆tTLG3months (both p=.03) but not for ΔtMTV6months and ΔtTLG6months (both p=.07).
Large variations in these parameters were observed for tTAD< 200cGy, while a more predictable response was observed for tTAD ≥200cGy. Note that negative values represent an increase in ΔtMTV and ΔtTLG. Signi cant differences annotated by asterisks Figure 7 a ∆tMTV3months b ∆tMTV6months c ∆tTLG3months and d ∆tTLG6months plotted against tTAD for the highand low lilotomab groups. A decrease in ∆tMTV or ∆tTLG is here shown as a positive percentage, and an increase correspondingly as a negative percentage. Statistically signi cant decreases in PET parameters were observed with increasing tTAD after 3 and 6 months in the high lilotomab group, but not in the low lilotomab group. It may be that the overall good response for low lilotomab group masks such a correlation. The results from the Spearman rank correlation tests are indicated for each group. Each symbol represents an individual patient Figure 8 a Absorbed dose to the total tumor volume, tTAD, in the four clinical response categories. Higher tTAD was observed in patients with CR, compared to SD and PD. b tTAD for response categories grouped as responders (CR+PR; in green) and non-responders (SD+PD; in red), and further strati ed by low and high lilotomab. Responders had a signi cantly higher tTAD than non-responders in the high lilotomab group. In the low lilotomab group only two patients were non-responders, and variations in tTAD were large for responders. Signi cant difference annotated by asterisks. c Responders and non-responders strati ed by a 200cGy threshold. All nonresponders had tTAD <200cGy, while all with tTAD≥200 were responders

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementarytable.docx