Serum methylmalonic acid concentrations at breast cancer diagnosis significantly correlate with clinical frailty

Abstract

Methylmalonic acid (MMA), a by-product of propionate metabolism, is known to increase with age. This study investigates the potential of serum MMA concentrations as a biomarker for age-related clinical frailty in older patients with breast cancer. One hundred nineteen patients ≥ 70 years old with early-stage breast cancer were included (median age 76 years). G8 screening, full geriatric assessment, clinical parameters (i.e., estimated glomerular filtration rate (eGFR) and body mass index (BMI)), and serum sample collection were collected at breast cancer diagnosis before any therapy was administered. MMA concentrations were measured via liquid chromatography with tandem mass spectrometry. MMA concentrations significantly increased with age and eGFR (all P < 0.001) in this older population. The group with an abnormal G8 (≤ 14, 51% of patients) had significantly higher MMA levels than the group with normal G8 (> 14, 49%): 260 nmol/L vs. 188 nmol/L, respectively (P = 0.0004), even after correcting for age and eGFR (P = 0.001). Furthermore, in the detailed assessment, MMA concentrations correlated most with mobility (Eastern Cooperative Oncology Group (ECOG) Performance Status and Activities of Daily Living (ADL) tools, all P ≤ 0.02), comorbidity (Charlson Comorbidity Index (CCI) tool, P = 0.005), and polypharmacy (P < 0.001), whereas no significant associations were noted for instrumental ADL (IADL), Mini-Mental State Examination (MMSE), Geriatric Depression Scale-15 (GDS15), Mini Nutritional Assessment-Short Form (MNA-SF), and pain (all P > 0.1). In addition, our results showed that higher MMA levels correlate with poor overall survival in breast cancer patients (P = 0.003). Elevated serum MMA concentrations at initial diagnosis are significantly associated, not only with age but also independently with clinical frailty, suggesting a possible influence of MMA on clinical frailty in older patients with early-stage breast cancer.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11357-023-00908-0.

Introduction

Methylmalonic acid (MMA) is a by-product of branched-chain amino acid and odd-chain fatty acid metabolism. It was first described as an inborn toxic metabolite of methylmalonic acidemia, caused by an inherited deficiency of the mitochondrial enzyme methylmalonyl-CoA mutase (MCM). This condition leads to multifactorial physiological function decline [1], including impaired mobility [2], declined cognition ability [3], and other life-threatening symptoms or complications. In fact, a rapid and significant accumulation of MMA, ranging from 5 µmol/L to 10 mmol/L [4, 5], commonly manifests as a fatal pediatric disease, with a median overall survival of 2 years [6, 7].

Recent studies have shown that in healthy adults, there is an age-related and mild accumulation of circulating MMA, increasing from 130 nmol/L in healthy young individuals (< 15 years) to 240 nmol/L in healthy older adults (> 65 years) [8, 9]. However, the impact of this chronic and mild accumulation on physiological function and health in aging individuals has yet to be explored.

Aging is a well-known risk factor for natural systemic function decline, leading to clinical frailty. Numerous studies have demonstrated the close relationship between frailty and aging, resulting in premature aging and negative health outcomes in older adults [10]. Breast cancer is the leading cancer worldwide, with around one-third of cases diagnosed after the age of 70 [11, 12]. Therefore, we aimed to investigate whether there is an association between MMA and age-related clinical frailty and survival in older individuals with breast cancer.

Patients and methods

Patient selection

The University Hospitals Leuven (UH Leuven) possesses a large blood biobank with an associated clinical database of all new breast cancer patients diagnosed and treated at the institute since 2003 (biobanking protocol approved by the UH Leuven ethics committee, S63773).

For the present study, it was decided to select a “frailty” cohort to study the correlation of MMA with clinical frailty. The G8 geriatric screening tool was used as an indicator of frailty. An overview of patient selection is shown in Fig. 1. In brief, it was decided to focus on patients with triple-negative breast cancer (TNBC) or luminal B-like breast cancer (LumB-like breast cancer) as these subtypes are associated with the highest medical need for additional prognostic biomarkers, compared to other subtypes. TNBC was defined by immunohistochemistry (IHC) and is characterized by the absence of estrogen receptor (ER) and progesterone receptor (PR) expression (< 1% of cells positive) and the absence of human epidermal growth factor receptor 2 (HER2) overexpression/amplification [13]. Tumors showing a 3 + score for HER2 on IHC while being negative on fluorescence in situ hybridization (FISH) were excluded. LumB-like breast cancer was defined by ER positivity (≥ 1% expression), HER2 negativity, and histological grade II or III (a surrogate marker for Ki67 as this was not available for patients diagnosed in earlier years). Only female patients with unilateral, invasive breast cancer of no specific type (NST) were allowed in this study, excluding the rarer breast cancer subtypes that often show very different biology.

Fig. 1.

Overview of patient selection. In brief, from 1st of January 2003 to 30th of December 2015, 415 patients aged ≥ 70 years had geriatric screening results (G8 tool) available. Patients with baseline serum sample and G8 not collected within 30 days after initial diagnosis were excluded, as well as patients with other histology than NST, with upfront metastatic disease, or with insufficient serum left in the biobank, resulting in a study population of 119 patients for the frailty cohort. Serum collection interval was defined as the time from BC diagnosis date to the serum collection date to ensure all serum samples were collected before any treatment administration. Abbreviations: BCs, breast cancers; IDC, invasive ductal carcinoma; LumB-like, luminal B-like breast cancer; NST, non-special type; TNBC, triple-negative breast cancer. Created with BioRender.com

In addition to that, patients for this frailty cohort were required to be 70 years or higher and have a G8 geriatric screening test available at breast cancer diagnosis. Since G8 was systematically performed in Leuven from 2012 to 2015 within the framework of a trial on geriatric screening in cancer patients [14], patients for the frailty cohort were selected from that time period. A geriatric assessment (GA) was further performed for this population. Patients with abnormal G8 (≤ 14) were further planned for full geriatric assessment, including functional status assessed by Eastern Cooperative Oncology Group (ECOG) Performance Status, Katz’s Activities of Daily Living (ADL), and Lawton’s Instrumental Activities of Daily Living (IADL), cognition by Mini-Mental State Examination (MMSE), depression by 15-item Geriatric Depression Scale (GDS15), nutrition by Mini Nutritional Assessment-Short Form (MNA-SF), comorbidity by Charlson Comorbidity Index (CCI), and polypharmacy. In this cohort, however, most patients with normal G8 (score > 14) [15] completed the full GA as well. Patients were only allowed to participate if the baseline blood sample and G8 had been collected within 30 days after initial diagnosis.

The current project was reviewed and approved by the local medical ethics committee of UH Leuven (S64752). The study was conducted according to EU legislation regarding ethical regulations and was registered online (NCT05352737, clinicaltrails.gov).

MMA and eGFR measuring

Serum samples were collected at the first consultation at the breast clinic, before the initiation of any treatment. All samples were coded, centrifuged (1600 × g for 15 min at 4 °C), aliquoted, and then stored at − 80 °C in the official UH Leuven biobank facility. MMA was measured using liquid chromatography with tandem mass spectrometry (LC–MS/MS), and data were processed by an external qualified laboratory (Algemeen Medisch Laboratorium, Antwerp, Belgium). In brief, master stock solutions of MMA and d3-MMA (1 mg/mL) were prepared in methanol/acetonitrile (80:20, v:v). Four individual MMA calibration solutions were prepared from the stock solution in water (170 to 1700 nmol/L) and stored at 2–8 °C until analysis. MMA extraction was performed on a Janus pipetting workstation (Perkin Elmer) at 4 °C from 50 µL serum with 130 µL of acetonitrile/methanol (80:20, v:v) extraction buffers containing d3-MMA internal standard at 1 µg/mL; the supernatants were then dried under nitrogen. Samples were reconstituted in 200 µL of mobile phase A (0.1% formic acid). After an injection of 10 µL of the sample, the separation of MMA from succinic acid and other compounds was performed on a Shimadzu LC system with a Kinetex 2.6 µm XB C18 100 × 4.6 mm column at a flow rate of 0.6 mL/min and a temperature set at 30 °C. A gradient was applied for 5 min (solvent A, 0.1% formic acid; solvent B, 0.1% formic acid in acetonitrile) to separate metabolites (0 min, 90%A; 1 min, 90%A; 3.5 min, 75%A; 4 min: 0%A; 4.5 min, 0%A; 5 min, 90%A). The LC system (LC30AD from Shimadzu) was coupled to tandem mass spectrometry (Qtrap 6500 + from Sciex) and operated in negative mode. Two transitions were analyzed for the MMA compound (116.914 > 73, collision energy (CE) 14 V and 116.914 > 55, CE 34 V), and one transition was analyzed for d3-MMA (119.879 > 76, CE 14 V) in multiple-reaction monitoring (MRM) mode. MultiQuant was used for data analysis.

As renal function is also known to influence MMA levels, serum creatinine (SCr) levels, measured at diagnosis date via an enzymatic assay with colorimetric detection on a Roche Cobas chemistry analyzer (Roche, Mannheim, Germany), were also retrieved from the patient’s medical record. The creatinine-based 2021 CKD-EPI formula was used to obtain the estimated glomerular filtration rate (eGFR)¹⁶.

Statistical analyses

For one patient in our study, an exceptionally high MMA level of 2724 nM was observed. We conducted a comprehensive review of the clinical information, considering factors such as age (85 years), diagnosis year (March 2013, indicating sample storage year as samples collected at the initial diagnosis), and eGFR (37.8 mL/min/1.73 m²). Despite being identified as a statistical outlier, no clear clinical explanation could be found for the remarkably elevated MMA level. Notably, other patients with similar age and eGFR showed lower MMA levels. For instance, a patient aged 89 with an eGFR of 18.6 mL/min/1.73 m² exhibited an MMA level of 371 nM, while a patient aged 70 with an eGFR of 25.8 mL/min/1.73 m² presented an MMA level of 730 nM. Considering these findings, we decided to include all patients in the final analysis, including the outlier.

In line, we conducted tests to assess the normality and log-normality of MMA level data. Although log-transforming improved the normality of the data, it did not entirely conform to a normal distribution. As a result, we utilized the Mann–Whitney U test to compare two groups (histologic grade and frailty) and the Kruskal–Wallis test for comparisons among three or more groups (tumor stage) to evaluate differences between groups. Furthermore, we employed linear regression, using log2 transformed MMA levels, to investigate the relationship between MMA levels and continuous variables such as age, eGFR, and BMI.

Analyses concerning the time to distant metastasis (distant metastasis-free survival (DMFS)) or death (overall survival (OS)) were conducted by modeling the data through Cox’s proportional hazards regression. January 1, 2021, was the latest censoring date. The prediction performance of models was computed (1) C-index, (2) net reclassification improvement by the continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI), and (3) net benefit through decision curve analysis.

All P values are two-sided, and statistical significance was defined as P < 0.05 unless otherwise specified. All statistical analyses were done with SAS software (v9.4; SAS Institute Inc., Cary, NC) and R software (v4.1.3).

Results

Baseline patient characteristics and MMA concentrations

The selection process of the cohorts is shown in Fig. 2, and the baseline characteristics of “non-frail” (G8 > 14) and “frail” (G8 ≤ 14) patients are represented in Table 1. In the entire frailty cohort, the median age was 76 years (interquartile range (IQR), 72–81), with a median MMA level of 219 nmol/L (IQR, 162–299) and median eGFR of 69 mL/min/1.73 m² (IQR, 55–86). About 51.2% of patients were classified as frail. Patients in the “frail” group were significantly older and had higher serum MMA levels compared with “non-frail” patients (all P ≤ 0.001).

Fig. 2.

Correlation between baseline serum MMA levels (log2 transformed) and clinical parameters in the frailty cohort. Linear regression was used for the scatterplot of age, eGFR, and BMI to calculate R² and P values. Mann–Whitney U test was used for the box plot of histologic grade and frailty, and Kruskal–Wallis test was used for the boxplot of tumor stage to calculate P values. Abbreviations: eGFR, estimated glomerular filtration rate; BMI, body mass index

Table 1.

Baseline characteristics of frail (G8 > 14/17) and non-frail (G8 ≤ 14/17) patients

	Non-frail, n = 61	Frail, n = 58	P value
Continuous, median (IQR)
Age, years	75 (71, 80)	79 (73, 83)	0.001
BMI, kg/m2	27 (25, 30)	27 (24, 30)	0.67
MMA, nmol/L	188 (147, 242)	260 (176, 354)	0.0004*
eGFR, mL/min/1.73 m2	73 (64, 87)	68 (53, 83)	0.20
Category, n (%)
Phenotype			0.82
LumB BC	55 (90)	53(91)
TNBC	6 (10)	5(9)
cTNM-stage			0.19
1	25 (41)	15 (26)
2	32 (52)	40 (69)
3	4 (7)	3 (5)
Histologic grade			0.995
II	40 (66)	38 (66)
III	21 (34)	20 (34)

IQR interquartile range

^*Comparison was conducted on log2 transformed data; Bold value indicates P < 0.05

Age and eGFR significantly correlate with serum MMA levels

Next, we sought to investigate the correlation between MMA and clinical parameters. From the univariate analyses displayed in Fig. 2, statistically significant heterogeneity in MMA levels was evident for persons of different ages and eGFR (all P ≤ 0.001), whereas BMI, tumor stage, and histologic grade apparently did not affect MMA levels (all P ≥ 0.17). A multivariate analysis was then performed including age and eGFR as potential covariates. Age and eGFR emerged as significant independent indicators of MMA levels (all P ≤ 0.03). For detailed information, see Supplementary material, Table S1.

Baseline MMA levels correlate with clinical frailty and overall survival

Given a significantly higher MMA level in the “frail” group compared to the “non-frail” group and a significant correlation between MMA and age as well as eGFR (Fig. 2), we next applied multivariate logistic regression. Noticeably, MMA independently correlated with G8 frailty also after adjusting for age and eGFR (P = 0.002). For detailed information, see Supplementary material, Table S2. Furthermore, we conducted principal component analysis (PCA) by transforming the three key parameters, namely, age, eGFR, and MMA, into principal components. Notably, MMA demonstrated a significant contribution to the principal components, as depicted in Figure S1a. Additionally, the first two principal components accounted for a substantial proportion of the total variance (84%), as illustrated in Figure S1b. Most importantly, both of these principal components exhibited significant associations with frailty (all P ≤ 0.02), as shown in Figure S1c. These findings further strengthen the robustness of our observation.

Next, we examined the potential associations of serum MMA levels with various geriatric assessment (GA) tools. Baseline full GA results are presented in Fig. 3a, providing a comprehensive overview of the assessments. Notably, Fig. 3b highlights significant correlations between MMA levels and the ECOG (rs = 0.25, P = 0.008), ADL (rs = 0.24, P = 0.02), CCI (rs = 0.26, P = 0.005), and polypharmacy (rs = 0.37, P < 0.001), whereas no significant associations were noted for IADL, MMSE, GDS15, MNA-SF, and pain (all P > 0.1) (see Fig. 3b).

Fig. 3.

Correlation between baseline serum MMA levels and clinical frailty and disease outcomes. a Bar plot of abnormalities in GA tools (in numbers and percentages). b Spearman rank correlation between MMA and GA tools in the frailty cohort. Asterisk denoted statistical significance for correlation. *P < 0.05, **P < 0.01, and ***P < 0.001. c, d Kaplan–Meier plots from the frailty cohort of baseline MMA levels in relation to DMFS (c) and OS (d); median MMA level was the cutoff value for high and low. e Decision curve analysis of long-term outcome (6-year overall survival) for the MMA-containing (purple curve) and clinical (blue line) risk prediction models. Abbreviations: ECOG, Eastern Cooperative Oncology Group Performance Status; (I)ADL, (Instrumental) Activities of Daily Living; MMSE, Mini-Mental State Examination; GDS15, Geriatric Depression Scale-15; MNA-SF, Mini Nutritional Assessment-Short Form; CCI, Charlson Comorbidity Index; DMFS, distant metastasis-free survival; OS, overall survival

In addition to its relationship with clinical frailty, we also explored the prognostic value of MMA. Kaplan–Meier analysis results of DMFS and OS are shown in Fig. 3c, d and clearly demonstrate the overall survival benefit of low baseline MMA (P = 0.003), while no such effect was noted for DMFS (P = 0.24). Similar results were obtained by applying Cox regression: higher MMA levels at diagnosis were not correlated with shorter DMFS (hazard ratio (HR) = 1.5, 95% CI 0.7 to 3.2), yet unfavorably affected overall survival (HR = 1.8, 95% CI 1.3 to 2.4). The latter association was also observed after correcting for covariates, including age, eGFR, and tumor stages, although it did not reach statistical significance (HR = 1.37, 95% CI 0.96 to 1.96).

The finding that MMA independently correlated with OS further led us to investigate whether MMA could perhaps add some prognostic value to the classic clinical prognostic markers of age and tumor stage. Compared with the model including only age and tumor stage as predictors, the MMA-containing model including age, tumor stage, and serum MMA level showed an increase in C statistic (i.e., from 0.78 to 0.81), as well as an improvement in reclassification as demonstrated by NRI of 0.36 (95% CI − 0.03 to 0.53) and IDI of 0.03 (95% CI 0 to 0.08). This improvement is quite remarkable, given that only MMA was added to the classical clinical model. Decision curve analysis showed that MMA had a better discriminative ability to recognize patients at high risk than the mere clinical parameter-based prognostic model (see Fig. 3e).

Discussion

MMA is known to increase with age in the general population [9, 17]. In the current study, we show for the first time that this correlation is also strongly present in patients with early breast cancer, indicating that the cancer process itself does not influence this relationship. Even more important, we show for the first time that chronic mild accumulation of MMA levels in the aged population also strongly and independently correlates with clinical frailty which reflects the functional decline accompanying chronological aging. Furthermore, our results indicate that MMA may be a promising prognostic factor for overall survival in breast cancer patients and reveal its consistent net benefit on the discrimination and reclassification ability on top of classical prognostic factors.

In this study, we show that MMA levels are significantly increased in frail older persons compared to non-frail older patients with breast cancer, even after correction for age and tumor stage. Studies attempted to reveal metabolomic markers for frailty [18–20], but MMA is mostly not included in metabolomics studies, possibly because of the extremely low abundance of MMA in human serum (nmol/L range). To our knowledge, our study provides the first evidence that elevated serum MMA is not only an aging marker but also an indicator of clinical frailty in older women with breast cancer. Specifically, our results indicate that chronic mild accumulation of MMA in older persons is associated with mobility decline, as witnessed by ECOG and ADL.

Although the nature of the study does not elucidate the underlying mechanisms of frailty, a previous study investigating the distribution of MMA after absorption in a mouse model has shown that skeletal muscle contains the highest concentration of MMA of all organs when the mice are alive and feeding [21]. The accumulation of MMA levels in tissue can induce cell damage by inhibiting respiratory chain complex II and the tricarboxylic acid cycle [22]. This has been confirmed in a patient diagnosed with methylmalonic acidemia (plasma MMA levels 45–147 µmol/L)²³, who presented with declined oxygen uptake and workload in an incremental exercise test. Histology from a biopsy of vastus lateralis muscle biopsy indicated widespread subsarcolemmal accumulation of mitochondria (SSAM) in trichrome and succinate dehydrogenase stains, which is the hallmark of a mitochondrial disorder, suggesting impaired energy metabolism. Although the MMA levels in that patient are still 1000-fold higher than in general older adults, the fact that skeletal muscle contains the highest MMA levels and that MMA can induce mitochondrial dysfunction may partially explain how the gradual accumulation of MMA levels can lead to mobility decline in older adults. However, a direct link between such low MMA concentrations and impaired skeletal muscle cell function needs further confirmation.

Notably, our study found that MMA is not correlated with cognitive decline in older adults with breast cancer, which differs from results seen in children with methylmalonic acidemia. High levels of MMA have been shown to lead to cerebral atrophy and neuro cell abnormalities, as demonstrated in a study where astroglia cells were exposed to high MMA levels (0.1–10 mmol/L) for 48 h [24]. This exposure compromised cell metabolic viability, oxygen consumption rate, glutamate uptake, and ATP content, indicating impaired energy metabolism and functionality of astroglia cells. MMA can also interfere with astroglia cell redox homeostasis and promote inflammatory responses [25], further perturbing their normal function. The inconsistency between our findings and published results may be partly explained by the relatively lower levels of MMA present in older adults, which are at the nmol/L range and 5000-fold lower compared to those in the aforementioned experiments or in children with methylmalonic acidemia. Such (moderately increased) levels of MMA in older persons may not be sufficient to cause severe cognitive damage. Furthermore, being an acid metabolite, high MMA concentrations can significantly alter the pH of the environment, with a concentration of 1 mmol/L causing the pH to drop to 2–3. Decreased pH has been shown to increase Aβ plaque load in a mouse model, resulting in Alzheimer’s disease [26]. It remains unknown whether the effects of extremely high MMA levels are due to pure MMA effects, or to an acidic environment caused by high MMA levels. Nevertheless, the absence of a correlation between gradual MMA levels and cognitive decline in older adults with breast cancer was further supported by a recent meta-analysis in the general population [27].

In our study, we found that MMA serves not only as an aging biomarker but also indicates age-related functional decline, known as clinical frailty. This finding is particularly important given that frailty is increasingly common among older patients and is associated with a 59–300% increased risk of death [28–30]. Although the extent and pattern of frailty vary among older individuals, with some experiencing declines in mobility, cognition, or other domains while others remain relatively healthy, our study was able to identify this heterogeneity by dissecting specific domains. Additionally, our results, with indirect supporting evidence from methylmalonic acidemia patients [21, 23], have the potential to develop therapy interventions to delay the aggressive frailty process in older persons. These findings provide a promising direction for further research on understanding and addressing frailty in older populations.

We realized that sample sizes were rather limited in our study; this was related to the fact that we aimed for very homogenous cohorts in terms of patient and tumor characteristics. To evaluate MMA levels in relation to frailty, it is mandatory to have access to both sufficient serum samples and extensive patient data including clinical frailty assessment, which is not routinely conducted in most clinical centers. Yet, we were able to show a clear link between MMA and frailty with statistical significance.

In conclusion, this study demonstrated for the first time that in a breast cancer patient population, baseline serum MMA levels at diagnosis are not only associated with chronological age but also with clinical frailty. Interestingly, unlike methylmalonic academia caused by inherited MCM deficiency, chronic progressive MMA accumulation with increasing age did not appear to correlate with impaired cognition, depression, or malnutrition but proved to be a strong indicator of mobility decline.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Conceptualization: Q.W., H.W., S-M.F., S.H., A.G. Provision of patient data and blood samples: H.W., K.P., P.N., A.S., I.N., C.K., G.F. Patient selection for current trial: Q.W., H.W., S-M.F., S.H. Methodology establishment: Q.W., C.K., J.B., H.W., S-M.F., S.H. Data analysis: Q.W., H.W., S-M.F., S.H. Writing of the manuscript: Q.W., H.W., S-M.F., and S.H. Critical review of manuscript: all authors.

Funding

HW is a recipient of the “Fonds voor Wetenschappelijk Onderzoek-Vlaanderen (FWO).” SMF acknowledges funding from the European Research Council under the ERC Consolidator Grant Agreement n. 771486–MetaRegulation, FWO Projects, Beug Foundation, Fonds Baillet Latour, KU Leuven FTBO/Internal Funding/CELSA, and Stichting tegen Kanker. The Gomes Lab is supported by a Pathway to Independence Award to A.P.G. from NCI (R00CA218686), a New Innovator Award from OD/NIH (DP2AG0776980), an ASC scholar award, Moffitt Cancer Center internal funds, and grants from the American Lung Association, the Florida Health Department Bankhead-Coley Research Program, the Florida Breast Cancer Foundation, the Phi Beta Psi Sorority, and METAvivor. PAM has received funding from Marie Curie Actions and the Beug Foundation, JFG is a FWO postdoctoral fellow, and AV is a recipient of a FWO PhD fellowship. The study was self-supported by other resources.

Data Availability

All data are available upon request from the authors.

Declarations

Conflict of interest

S-MF has received funding from Bayer AG, Merck, Black Belt Therapeutics, Gilead, and Alesta Therapeutics, has consulted for Fund + , and is in the advisory board of Alesta Therapeutics. KP: research grants to institute: MSD and Sanofi. Speaker fees and honoraria for consultancy and advisory board functions: Astra Zeneca, Eli Lilly, Exact Sciences, Focus Patient, Gilead, MSD, Novartis, Pfizer, Roche, Seagen. Speaker fees and honoraria for consultancy and advisory board functions to institution: Astra Zeneca, Eli Lilly, Exact Sciences, Gilead, MSD, Novartis, Pfizer, Roche, Seagen. Stock options: Need Inc. Travel grants from Astra Zeneca, Novartis, Pfizer, PharmaMar, Roche. Other authors declare that they have no competing interests.