Correlation of Baseline Tumor Burden with Clinical Outcome in Melanoma Patients Treated with Ipilimumab

Abstract Introduction: Tumor burden is a frequently mentioned parameter; however, a commonly accepted definition is still lacking. Methods: In this double-center prospective and retrospective study, 76 patients with unresectable stage III or stage IV melanoma treated with ipilimumab were included. We defined the baseline tumor burden (BTB) as the global sum of all metastases’ longest diameters before treatment started and correlated the calculated BTB with disease control rate (DCR), progression-free survival (PFS), overall survival (OS), and with the baseline levels of LDH, S100B, and sULPB2. Results: BTB correlated significantly with DCR (p = 0.009), PFS (p = 0.002), OS (p = 0.032), and the occurrence of NRAS mutation (p = 0.006). BTB was also correlated to baseline serum levels of LDH (p = 0.011), S100B (p = 0.027), and SULBP (p < 0.0001). Multivariate analysis revealed that BPB and LDH were independently correlated with PFS and OS. With increasing BTB, disease control was less likely; no patient with a BTB >200 mm achieved disease control. For patients with brain metastasis, no correlation of BTB with DCR (p = 0.251), PFS (p = 0.059), or OS (p = 0.981) was observed. Conclusion: Calculated BTB is an independent prognostic factor for patients with metastatic melanoma treated with ipilimumab. Using calculated BTB as a definition of tumor burden may help increase comparability of outcome of therapies in future studies.


Introduction
In the last decade, a variety of different therapeutic options for metastatic melanoma have been established.Ipilimumab was the first treatment to show a survival benefit for treated patients suffering from metastatic melanoma [1].Over the next years, BRAF and MEK inhibitors, more immune checkpoint blockades (ICB) such as PD1-blockade, and the immune combination therapy with nivolumab and ipilimumab have been added as treatments [2][3][4].However, the majority of patients will still suffer from progression and die eventually from metastatic disease.Several prognostic markers have been investigated: Eastern Cooperative Oncology Group performance status, serum level of S100B [5], or serum level of lactate dehydrogenase (LDH) have been primarily associated with the overall prognosis [6][7][8][9].Subsequently, tumor mutational load, pre-treatment metastatic growth rate, immune infiltration, ctDNA levels at baseline, and chemokine signatures, among others, have been shown to correlate with treatment outcome to ICB but have not been clinically validated as definite selection parameters to decide on the optimal first-line therapy [10][11][12][13][14].
Soluble natural killer group 2 member D (NKG2D) ligands consist of MHC-class-I-related chain A/B (MICA/B) or UL16-binding proteins (sULBP2).NKG2D is an activating receptor on all murine and human NK cells and is constitutively expressed on human CD8+ T cells [15].The ligands are normally absent on healthy cells and are highly expressed in tumor or virally infected cells [15][16][17].Several tumors including melanoma are able to release soluble NKG2D ligands to impair the tumor immunity.Accordingly, a high serum level of soluble NKG2D such as soluble ULBP2 has been identified as an indicator of poor prognosis and was even superior to S100B in this regard [11].
Tumor burden is used as a further parameter to decide on treatment approach; a high tumor burden demands usually a more aggressive treatment, since it has been associated with a worse outcome.However, despite being a widely utilized term, "tumor burden" has no generally accepted definition: it has been described in different publications as the number of metastatic organs or it has been equated with high levels of LDH or S100B [18][19][20].Tumor burden as calculated sum of the diameters of metastases has been proposed as an objective value and was correlated with the response rate to pembrolizumab in patients with advanced melanoma [19,21], but so far no analysis regarding the correlation of calculated tumor burden with ipilimumab or the immune combination therapy has been published.
In this prospective/retrospective study, we investigated whether the baseline tumor burden (BTB) defined as a sum of the lesions' longest diameters in patients with unresectable/metastatic melanoma treated with ipilimumab was predictive for the clinical outcome.Furthermore, we assessed the correlation between BTB and other potentially predictive markers such as LDH, S100B, and sULBP2.

Patient Collective
In this double-center prospective and retrospective study, 76 patients suffering from unresectable stage III or IV melanoma treated with ipilimumab between September 2012 and July 2014 were included, data cut off was May 2016.Patients were treated with the approved dose and schedule for ipilimumab of 3 mg/kg infused over 90 min, administered every 3 weeks for up to four doses.Adverse events (AEs) were graded using the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0.Clinical response was assessed with computed tomography (CT), positron-emission-tomography-CT (PET-CT), or magnetic resonance imaging (MRI).The baseline scan was obtained within 30 days before treatment started and all patients underwent initial post-therapy imaging.All procedures performed in this study were in accordance with the 1964 Helsinki Declaration and its later amendments.All prospectively included patients signed a written informed consent.All retrospectively included patients were irreversibly anonymized and the informed consent for these patients was waived.The study was approved by the Ethic Committee of the Ludwig-Maximilian-University, Munich, Germany.The number of the approval was 420-11.

Serum Analyses
The baseline serum level of LDH was assessed for all patients and of S100B for 33 patients in the laboratory of the clinical chemistry at the LMU Munich or at the University of Tübingen.

Analysis of ULBP2 from Sera
Blood samples from 21 patients were analyzed at baseline.All samples were obtained and processed following a standardized protocol.Briefly, venous blood was drawn into gel-coated serum tubes (Sarstedt), clotted at room temperature for 30-60 min, and thereafter centrifuged at 2,000 g for 20 min.Serum was extracted and immediately frozen at −80°C.All samples were thawed in the morning 1 h before the analysis, and patients' sera were analyzed for sNKG2DL by ELISA.Determination of sULBP2 was carried out with the mouse anti-ULBP2 capture mAb (R&D Systems, Minneapolis, MN, USA) and the biotinylated goat anti-ULBP2 detection mAb (R&D Systems).On the previous evening, the analysis plates were coated with the capture anti-ULBP2 mAb at 4.0 μg/mL in PBS.In the morning, they were blocked by addition of 300 μL of 1% bovine serum albumin (BSA) in PBS for 1 h at room temperature.After washing 100 µL recombinant human ULBP2 diluted with 1% BSA in PBS to a high standard concentration of 500 pg/mL (R&D Systems), the patient's samples were added, covered with an adhesive strip, and incubated for 2 h at room temperature.Sera were not diluted before being added to the plates.After incubation and washing, the detection biotinylated goat anti-human ULBP2 mAb (R&D Systems) at 200 ng/mL in 1% BSA in PBS was added for 2 h at room temperature.After washing, streptavidin-HRP (1:200 in 1% BSA/PBS) was added for 20 min at room temperature.After washing, plates were developed using a mixture of tetramethylbenzidine and hydrogen peroxide (1:1, R&D Systems, Minneapolis, MN, USA).Immediately after 50 µL of 2 N sulfuric acid were added to each well, the absorbance was measured at 450 nm with a wavelength correction of 540 nm.Results are indicated as means of triplicates.

BTB Assessment
The assessment of BTB was based on the CT, MRI, or PET-CT measurements conducted by a radiologist before treatment started with ipilimumab.BTB was defined as the global sum of all metastases' longest diameters in a patient.For lesions, which were not exactly measured but unequivocally identified as metastases, we Baseline Tumor Burden as Prognostic Factor for Outcome Oncology 2024;102:76-84 DOI: 10.1159/000533504 had to determine a diameter that would fit to the pathology of a metastasis and would not incorrectly extend or diminish the BTB.The following values were determined: for every metastatic lymph node or soft tissue lesion with uncertain diameter, the size of 15 mm was being used.In the case of (sub)cutaneous, hepatic, or osseous metastases, the size of 10 mm diameter was applied.For multiple disseminated lung metastases (>5 lesions), a value of 60 mm was assumed, and if the pulmonal metastases were described as multiple micrometastases, a sum of 30 mm diameter was assumed.A summarized diameter of 20 mm was applied for peritoneal or pleural carcinomatosis, presented as a circulated nodular increase of contrast density.

Disease Control Rate
The response of the first staging after the last therapy cycles of ipilimumab was evaluated and graded by radiologists.Patients with complete response, partial response, and stable disease were defined as patients with controlled disease, and patients with mixed response or progression were defined as patients with progressive disease.

Statistics
Statistical analysis was performed using SPSS (IBM SPSS Statistics version 24; IBM Corporation, Armonk, NY, USA) and GraphPad Prism software (version 5.01; GraphPad Software Inc., San Diego, CA, USA).Analysis of differences between two normally distributed test groups was performed using the Student's t test.Welch's correction was applied to Student's t test datasets with significant differences in variance.For analysis of 3 or more groups, the one-way ANOVA test was performed without a posttest.Overall survival (OS) of the patients was calculated from the time span of first administration of ipilimumab treatment until death or last contact to patient.Correlation was calculated using Spearman rho correlations test.χ 2 test and multivariate analysis were performed to determine the independency of the variables.Survival curves and progression fractions were analyzed with the product limit (Kaplan-Meier) method for censored failure time data assuming proportional hazards.Univariate comparisons of Kaplan-Meier estimators were performed with the log-rank test.Two-tailed p values were calculated.A p value of <0.05 was considered significant with *p < 0.05, **p < 0.01, ***p < 0.001.

Study Population
Seventy-six patients who were treated with ipilimumab between September 2012 and July 2014 were included in the study.We recruited 22 patients right before treatment with ipilimumab and 54 patients retrospectively after treatment with ipilimumab has been concluded.The median age of the patients was 65 years (30-85 years).53% of the patients were female.Most patients (89%) suffered from stage IV melanoma and 11% from unresectable stage III melanoma.Nine (12%) patients had brain metastasis.The time between the primary diagnosis until progression was significantly correlated with the OS in stage IV or unresectable stage III: patients progressing in <12 months after the primary diagnosis survived shorter time after treatment with ipilimumab than patients who progressed >12 months after the primary diagnosis (p = 0.006).Twenty-one (28%) of the patients carried a BRAF V600 mutation, and in 13 (17%) of the patients, a NRAS Q61 mutation was detected.Only 21% of the patients received ipilimumab as first-line therapy, most patients (41%) were treated with ipilimumab as third or even higher line therapy.All patients with BRAF V600 mutation received a BRAF inhibitor previously; other previous treatments included chemotherapies (53%) or MEK inhibitors (13%), either pimasertib or trametinib.Interestingly, there was no correlation with progression-free survival (PFS) (p = 0.722) nor OS (p = 0.359) and the line of ipilimumab therapy.The majority (76%) of the patients received 4 cycles of treatment with ipilimumab, and 7% were treated with three cycles, 11% with two cycles, and 7% with only one cycle.All patients with controlled disease received all 4 treatment cycles.The number of the applied treatment cycles were significantly correlated to PFS (p < 0.001) and OS (p = 0.006).Treatment discontinuation was either due to progression-related death (39%) or due to severe AEs.The most frequent AE ≥ grade 2 according to CTCAE 4.0 were autoimmune colitis (19%), hypophysitis (11%), and exanthema (9%).Nephritis, hepatitis, coxitis, and elevated pancreatic enzyme levels were also observed.
The overall response rate was low, only 11% of the patients responded to treatment, the disease control rate (DCR) comprised 24% of the patients with 6 complete responders, 2 partial responders, and 10 patients with stable disease.At the time of data analysis, 11 (15%) patients were lost to follow-up and 53 (70%) patients had died.The median follow-up time for OS conducted was 29.5 months.

Assessment of BTB
As previously described, tumor burden is a widely used but not clearly defined term in oncology.We defined BTB as the sum of largest diameter of the metastases in MRI, PET-CT, or CT scan results before treatment.BTB ranged from 15 mm up to 426 mm; the median BTB was 107 mm.BTB was significantly correlated with DCR (p = 0.009), with PFS (p = 0.002), and with OS (p = 0.032) (shown in Fig. 1a-c).We also observed an upper limit of BTB for disease control: patients with a BTB of >200 mm never achieved disease control after ipilimumab treatment in any case.
BTB was not correlated with line of treatment of ipilimumab (p = 0.571) nor the occurrence of BRAF mutations (p = 0.498), but with the prevalence of NRAS Angelova-Toshkina/Weide/Tietze/Hebst/ Tietze mutations (p = 0.006) (shown in Fig. 2a).BTB was also significantly correlated to the number of metastatic organs (shown in Fig. 2b) (p < 0.0001), but no significant correlation with BTB and the occurrence of liver metastasis was observed (p = 0.072).

Analysis of Correlation of Brain Metastasis, BTB, and PFS
In our cohort, 9 patients suffered from brain metastasis.The occurrence of brain metastasis is a well-defined independent parameter for a worse outcome for patients with melanoma [8,22].We were curious, whether brain metastasis is more relevant for the outcome than BTB, and indeed, we did not see a significant correlation of DCR (p = 0.251), PFS (p = 0.059), or OS (p = 0.981) and BTB in patients suffering from brain metastasis (shown in Fig. 3a, b).After exclusion of patients with brain metastasis, the correlation of PFS and BTB was even more pronounced (p < 0.001).Based on these data on admittedly few patients, the occurrence of brain metastasis seems to be an independent and even more relevant prognostic factor regarding the treatment outcome to ipilimumab than BTB.

Assessment of LDH, S100B, sULPB2
Elevated serum levels of LDH and S100B have been shown to be associated with a worse prognosis [8,19,23,24].We were able to confirm the significant inverse correlation of baseline level of LDH with DCR (p = 0.008), PFS (p = 0.002), and OS (p = 0.003) as well as the significant inverse correlation of baseline level of S100B with DCR (p = 0.016), PFS (p < 0.001), and OS (p = 0.008) in our analysis.We next assessed the NKG2D ligand sULBP2, since the occurrence of this soluble ligand has also been correlated with tumor burden and clinical outcome in patients with metastatic melanoma [10,11].The average concentration of this protein in our patients was 167.04 pg/mL.The baseline level of sULBP2 among the patients with controlled disease had a median value of 7.81 pg/mL and the baseline level of sULBP in nonresponders was 235.27 pg/mL much higher than in responders, the difference was significant (p = 0.012).In contrary to the published data, no significant correlation of the baseline level of sULBP with PFS (p = 0.057) or OS (p = 0.216) could be shown.

Correlation of sULPB2, LDH, S100B, and BTB
The baseline level of sULBP2 correlated highly significantly with BTB (p < 0.0001) (shown in Fig. 4a).The baseline levels of S100B and more frequently of LDH have been used as an equivalent for tumor burden in several studies [25][26][27].We were able to observe a significant correlation of BTB with the baseline level of LDH (p = 0.011) and S100B (p = 0.027) (shown in Fig. 4b, c); in our patients, however, the correlation of the baseline levels of LDH and S100B was not highly significant.It was therefore in question, whether they were independent or dependent predictive factors.Multivariate analysis revealed that the BTB was an independent marker from LDH for PFS and OS.The multivariate analysis for BTB and S100B was not conclusive.

Characterizing Patient Groups with Low, Moderate, or High BTB
The median BTB was 107 mm.All patients with a BTB above 200 mm progressed after ipilimumab treatment (shown in Fig. 1a).Based on this observation, the patients were grouped into three groups with low (up to 100 mm, 46%), moderate (101-200 mm, 40%), or high (>200 mm, 14%) BTB (shown in Fig. 5a).31% of the patients with Angelova-Toshkina/Weide/Tietze/Hebst/ Tietze low, 23% with medium, and no patient with high BTB experienced controlled disease after treatment.Demographic data did not differ between the groups (Table 1).The correlation between PFS and BTB was greater in analyzing the PFS of the 3 groups with low, moderate, and high BTB (p = 0.001) (shown in Fig. 5b) and patients with a low BTB survived significantly longer (p = 0.003) (shown in Fig. 5c).

Discussion
To our knowledge, this is the first study to assess the prognostic effect of BTB on clinical outcome of patients with metastatic melanoma treated with ipilimumab.Tumor burden has been associated with multiple metastatic organs or elevated levels of LDH, but no consistent definition of this term exists so far, even though its assessment is very relevant in the clinical routine.Only lately the sum of diameters of the metastases before treatment start has been used as a definition of tumor burden for melanoma treated with pembrolizumab and melanoma and small lung cancer treated with ICB [19,28].
We showed that BTB was correlated with DCR, PFS, and OS after treatment with ipilimumab.Not 1 patient with a BTB higher than 200 mm responded to treatment with ipilimumab, which is consistent with Joseph et al., who showed that higher BTB was associated with decreased response also in anti-PD1-treated melanoma patients as well [19].In our study, NRAS-mutated patients suffered from a higher BTB than the WT patients, and indeed NRAS mutated melanoma has been described to be more aggressive [29].SULBP2 and S100B have been shown to be strong prognostic indicators [8,11,19,23,24].The release of NKG2D ligands, such as sULBP2, promotes systemic downregulation, and dysfunction of the NKG2D receptor on lymphocytes [30].We observed a correlation with the levels of sULBP2 and BTB, which is consistent with published data: an increase of the concentration of sULBP2 has been correlated with a higher tumor load; however, tumor load was defined in this publication as measurable disease compared to clinically non-apparent tumor manifestations, but not as in our study as differences in the mass of the metastases [11].
Elevated serum levels of LDH have been associated with a worse prognosis, which is consistent with our data, but an association with higher tumor burden has been controversially discussed [8,19,23,24].Nevertheless, LDH levels have been frequently used as a substitute for high tumor burden in clinical studies [26,27,[31][32][33].According to our data, LDH does not seem to be a surrogate for BTB: both variables were merely but not strongly correlated; we observed patients with high LDH and low BTB and patients with high BTB and low LDH and the multivariate analysis confirmed that both values were independently associated with PFS and OS.This is consistent with a recent review, which highlights the complexity of LDH [34].LDH is a biomarker associated with the activation of several oncogenic signaling pathways as well as with the metabolic activity invasiveness [34].It is unfortunate that key opinion studies such as the DREAMseq Trial, in which the most efficient sequence of BRAF/MEK inhibition and ICB combination therapy has been evaluated, also only used LDH as a marker for tumor burden but did not assess the calculated BTB [32].Large tumors have been shown to be increasingly immunosuppressive since they impair the ICB induced response of the host immune system [35].It would be therefore crucial to acquire data, which specifically address the question of whether patients with high calculated BTB would still benefit more from ICB as first line.Furthermore, the European consensus guideline evaluates debulking procedures critically, as there is no evidence of improved survival [36].However, considering the association of BTB with ORR, PFS, and OS in ipilimumab-and pembrolizumab-treated patients, it remains in question, whether reduction of tumor burden before initiation of ICB may be useful in some scenarios, which could be addressed in further studies.
This study has several limitations, such as the relatively small patient cohort and its duo-centric prospective/ retrospective design.Another limitation is the nonexisting gold standard for BTB calculation.We included in the BTB calculation all metastases, what in some cases lead to an approximation in the calculation, as depicted in the methods.
BTB is an easily calculated independent prognostic factor for clinical outcome in patients with metastatic melanoma treated with ipilimumab.Since tumor burden is relevant for response to most therapies, using calculated BTB as a definition of tumor burden may help increase comparability of the outcome of therapies in future studies.

Statement of Ethics
The study was approved by the Ethic Committee of the Ludwig-Maximilian-University, Munich, Germany.The number of the approval was 420-11.All procedures performed in this study were in

Table 1 .
Comparison of the three groups with low, moderate, and high BTB regarding demographic data, affected organ systems, and baseline serum levels of LDH, S100B, and sULPB2 Toshkina/Weide/Tietze/Hebst/ Tietze accordance with the 1964 Helsinki Declaration and its later amendments.All prospectively included patients signed a written informed consent.Informed consent for retrospectively included patients was waived because the patients were irreversibly anonymized.