Variability in resistance training trajectories of breast cancer patients undergoing therapy

Variability in resistance training trajectories of breast cancer patients undergoing therapy | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Variability in resistance training trajectories of breast cancer patients undergoing therapy Maximilian Koeppel, Karen Steindorf, Martina Schmidt, Friederike Rosenberger, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4089501/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 10 Dec, 2024 Read the published version in Supportive Care in Cancer → Version 1 posted 13 You are reading this latest preprint version Abstract Purpose In resistance training (RT) the change in training volume from training sessions (TS) to TS, is an indicator of training progress. Resulting growth-trajectories are likely to differ between individuals. Understanding this variation is important for exercise planning in general, but even more for clinical populations. We investigated this variation in breast cancer patients undergoing treatment. Methods Data of 69 patients from two randomized controlled trails were investigated. They conducted a 12-week RT program. We fitted a quadratic Bayesian regression model to the baseline standardized training volume over the course of the intervention. We allowed all parameters to vary both between exercises and between individuals. Results We observed a positive linear component of 0.093 (95% Uncertainty interval (UI) 0.058 to 0.120) and a negative quadratic component of -0.002 (95% UI -0.008 to 0.001) for the mean trajectory of the change in training volume. For the different exercises we observed a dispersion for both the linear (0.043, 95% UI 0.018 to 0.082) and the quadratic component (0.002, 95% UI < 0.001 to 0.004). Variation between-individual appears to be approximately 4 times larger. We also observed between-exercise variation within individuals. Extrapolation of the regression model indicates training progression stagnates after 20.6 TS (95% UI 14.8 to 44.4). Conclusion There is substantial variation in RT response between breast cancer patients undergoing tumor therapy and in-between exercises. The non-linear trajectory indicates that training progression will eventually plateau, demanding periodization and timely modification. Trial Registration: BEATE Study: NCT01106820, Date: April 20, 2010; BEST Study: NCT01468766, Date: November 9, 2011 exercise oncology adjuvant tumor treatment response variability hierarchical model Bayesian statistics Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Based on a large body of evidence, resistance training provides several positive effects for cancer patients. However, RT effects do not follow a simple causal stimulus-response relationship between mechanical input and physiological adaptation, but are subject to a complex network of effect modificators [ 16 , 40 ]. Thus, RT effects may vary between individuals and leads to the classification of individuals into distinct response-categories [ 1 , 3 , 10 , 13 , 17 , 25 ]. In contrast to medical oncology, in which response refers to the efficacy of tumor treatment in reducing tumor size or severity [ 27 ] the in RT research the term is used more ambiguously and could refer to several RT related outcomes of interest, such as the one repetition maximum (1RM), the cross sectional area of a particular muscle [ 17 ], or the performance in a functional test [ 10 ]. This observations has stimulated scientific endeavors over the last few years and gave rise to controversial discussions. One major criticism is that studies are often confined to comparing data of only two distinct time points with each other. Where the first point is timed before the start of the exercise intervention and the second one right after the end of the training period. Thus, the two measurements are prone to within-individual variation, such as different mood, motivation and pain tolerance, but also measurement error which can lead to misestimations [ 2 , 16 ]. Another statistical artifact, would be that extreme values in the first assessment tend to be drawn closer to the mean in the second assessment, resulting in an overestimation of the effect variability [ 4 ]. From a more practical perspective, it is well known that strength gains follow a non-linear time-trajectory, with the steepest incline at the beginning of the intervention, which eventually approaches a plateau during the course of the intervention [ 28 , 32 ]. By ignoring the qualitative differences, linear models tend to underestimate the strength gains in the early stages of training and overestimate the later ones. Furthermore, strength gains do not appear systemic but locally, involving primarily those muscles engaged in the particular exercise [ 26 ]. Therefore, it is surprising that the RT response variability within individuals has not received much attention so far. Although response variability to exercise has been mentioned in several well placed publications, in exercise oncology [ 11 , 18 , 36 ] it did not attract the researchers’ focus of interest. However, the investigation of exercise response variability is especially important for clinical populations, to ensure the optimal care of these populations. Ignoring the effect variability could lead to false expectations for the individual patient, due to an overconfidence in a one-size-fits-all approach [ 21 ]. It is the purpose of this study to investigate how the change in training volume varies between exercises and between individuals undergoing adjuvant cancer treatment over the course of an RT intervention. In addition, we are interested if, and at what point in time these training-trajectories might reach a plateau. Due to the nested structure of the data we fitted a hierarchical Bayesian regression model, which allowed us to estimate the variability between exercises and between individuals. In order to model a potentially non-linear time-volume relationship and estimate the number of TS leading to a plateau, we included a quadratic term to the model. Method Design and participants We conducted a secondary analysis by pooling the data of the BEATE (NCT01106820) and the BEST trial (NCT01468766), two randomized clinical trials conducted with physically inactive ( 18 years) undergoing adjuvant chemotherapy (BEATE trial) or adjuvant radiation therapy (BEST trial). Detailed information about the study design are described elsewhere [ 29 , 34 ]. Briefly summarized, both trials investigated the effect of a 12-week machine-based RT intervention on cancer-related fatigue. These studies were approved by the ethics committee (S-012/2009, S-447/2010), were conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all patients. A total of 261 patients (101 BEATE, 160 BEST) were enrolled in both trials with 132 assigned to the intervention group (52 BEATE, 80 BEST). While many patients trained at the university’s training facility (41 BEATE, 31 BEST), the patients were also given the option to train in facilities near their homes in order to reduce the barriers for participation. For this analysis, we only included the 69 patients who trained at the study center’s facility to reduce additional sources of variance (Fig. 1). Baseline patient characteristics are outlined in Table 1 . Intervention In accordance with the ACSM exercise guidelines for cancer survivors, current during the conduction of the study [ 35 ], the training was planned to take place twice a week over the course of 12 weeks, resulting in a maximum number of 24 TS. In both studies the same core exercises were conducted: leg press, leg extension, leg curl, shoulder internal and external rotation, seated row, latissimus pull down and butterfly. Patients who participated in the BEATE trial also performed butterfly reverse each TS and anteversion or retroversion of the shoulder every other TS. Each exercise was conducted for 1 to 3 sets with 12 repetitions at 60–80% of their hypothetical one repetition maximum, estimated with the Brzycki-Formula [ 7 ]. The training schedule followed a progressive approach, in which the applied load was increased by at least 5% if the prescribed load was successfully lifted for 3 sets of 12 repetitions in three consecutive TS. The TS were supervised by experienced exercise- or physical therapists. For more details please refer to the design [ 29 , 34 ] and primary endpoint papers [ 33 , 37 ] as well as the CERT-Checklist (Supplementary Information SI 9). Construction of outcome and predictor variables The goal of the analysis was to investigate the change in training volume over time. Therefore the number of the TS was chosen as predictor variable. We excluded the first two TSs from each patient since those were considered familiarization sessions. Training volume per TS (product of load, number of sets and number of repetitions realized in the particular session) was chosen as our best available proxy in determining the training progression. The training volume per TS for each exercise was z-standardized to the baseline mean and baseline standard deviation of the particular exercise and served as outcome variable for all further analyses. Thus, the resulting regression coefficients can be interpreted and will be referred to as standardized mean differences (SMD). TS were considered as valid if more than eight repetitions and more than two sets were performed. If the number of repetitions exceeded 12, training volume was equalized with the volume of the next valid session to avoid misestimations. Management of outliers is discussed in (Supplementary Information SI 1). Statistical analysis We applied a quadratic hierarchical Bayesian regression model with training volume as outcome variable and number of TS as predictor variable. We added a quadratic term to the model, assuming that training progression does not follow a linear path but will display a steep incline in the first training sessions and eventually approaching a plateau [ 28 , 32 ]. The model consists of three levels: The individual TS (Level 1) which are assigned to the different exercises (Level 2), which are then assigned to the patients (Level 3). This model allowed all parameters (i.e. intercept, linear component and quadratic component) to vary on the 2nd and 3rd level and enabled us to investigate the variability of the model parameters between exercises (Level 2), between individuals (Level 3) but also the between exercises within individuals. Regarding the variability of the RT response trajectories, we focus on the linear component, i.e. a positive linear component describes a positive response, and a negative linear component describes an adverse response. We chose a Bayesian framework since it allowed us to integrate a large number of already existing information into the prior distribution. Precisely we used the data of an unpublished meta-analysis of 25 resistance training studies with breast cancer patients and survivors, including a total of 112 effect sizes (Supplementary Information SI 3). Furthermore the Bayesian analysis does not result in a point estimate for each parameter but in a probability distribution, which allowed us probabilistic interpretations of the parameter estimates[ 41 ]. To summarize the posterior distribution of a parameter, we chose the posterior distribution’s mean as the most probable location of the parameter and its 95% uncertainty interval (95% UI), thus, describing the range of the parameter values where the true value can be expected with a 95% probability. For simplicity’s sake, we will be referring to the mean of the posterior distribution as the parameter estimate if not otherwise specified. The analysis was conducted via in R with the brms package[ 8 ]. 95% Heterogeneity Intervals (95% HI) In accordance with Bolger et al. [ 6 ] we also calculated the 95% HI for the exercises ( \({95\%HI}_{Exercise}\) ) and the full model ( \({95\%HI}_{Full}\) ). The \({95\%HI}_{Exercise}\) states (under the assumption of normality) that the parameter estimate of any exercise for this specific population is expected to lie within this interval with a probability of 95%. For clarification, this differs from the definition of the 95% UI, since the 95% UI describes where the parameter estimate of one exercise can be expected with a 95% probability. Analogously, \({95\%HI}_{Full}\) , refers to the range of values where any parameter estimate, regardless of person or exercise, would be expected with a probability of 95%. Training Sessions until Plateau (TS Plateau ) The first derivative of the quadratic regression formula allows to calculate the vertex of the function: $${TS}_{Plateau}=\frac{LinearComponent}{2*QuadraticComponent}$$ This vertex serves as the indicator, where the trajectory will reach its plateau. To model this vertex, 10,000 values were randomly drawn from the posterior distributions of the linear and the quadratic component, resulting in 10,000 pairs of estimates. Each of these pairs was entered into the formula introduced above. This in turn resulted in a probability distribution consisting of 10,000 values for \({TS}_{Plateau}\) . In case of profoundly skewed posterior distributions, we chose to report the median of the posterior distribution. To ensure the transparency and reliability of the analysis, we followed the WAMBS-Checklist [ 12 ] (Supplementary Information SI 4 for the analysis with diffuse priors). In addition, we fitted linear two-level-models to the data of each exercise and compared them to their quadratic counterparts, using leave-one-out cross validation (Supplementary Information SI 5) to justify the utilization of the quadratic model. Finally, we reran the three-level-model with only those individuals attending at least 50% of TSs (n = 38, 53%) and the ones attending at least 75% (n = 27, 38%) of TS (Fig. 1) to check if the parameter estimates are skewed by systematically low attendance (Supplementary Information SI 6). Results The raw baseline values for all exercises are provided in Table 1 . Summary statistics, such as the means, standard deviations (SD) and 95% UI for the posterior distributions from the main analysis are displayed in Table 2 . A comparison between the linear and the quadratic model showed a superiority of the quadratic model (Supplementary Information SI 5), hence, only the results from the quadratic model will be displayed in the main manuscript. Main Analysis The main analysis yielded an average linear component of SMD = 0.093 (95%UI: 0.058 to 0.120) per TS and an average quadratic component of SMD= -0.002 (95%UI: -0.008 to -0.001) per TS-squared. Variability between Exercises Between exercises we observed a standard deviation of SD = 0.043 (95%UI: 0.018 to 0.082) for the linear component and SD = 0.002 (95%UI: <0.001 to 0.004) for the quadratic component. Figure 2 and Table 3 show that the posterior distributions of the linear components for all exercises are right of the zero-line, indicating positive training response across exercises. This is also supported by the 95%HI Exercise , ranging from 0.013 SMD (95%UI: -0.094 to 0.077) to 0.175 SMD (95%UI: 0.140 to 0.234) and a median coefficient of variation at about 45% (95% UI = 19 to 88%). The 95%HI Exercise of the quadratic component ranges from − 0.006 SMD (95% UI = -0.009 to -0.003) to 0.001 SMD (95% UI = -0.002 to 0.007). This means that almost any combination of both components results in an inversely L-shaped curve. The variation between exercises also has an impact on the proportion of individuals who display a positive response to a particular exercise, i.e. the proportion of individuals whose posterior distribution’s median is positive, ranging from 77% for the rowing exercise to 98% for the butterfly exercise (Supplementary Information SI 7 & 8). Variability between Individuals Between individuals the linear component dispersed with a standard deviation of SD = 0.155 (95% UI: 0.142 to 0.168) and the quadratic component with SD = 0.008 (95% UI: 0.007 to 0.009). The variability of the linear component between individuals appears to be 3.6 times the size of the variation between exercises (95% UI: 3.3 to 3.9). Concerning the quadratic component, the inter-individual variability appears to be four times larger than the variation observed in the quadratic components for the different exercises (95% UI: 3.5 to 4.5). Considering the total variation, i.e. individual variation across exercises, the 95%HI Full for the linear component ranges from − 0.290 SMD (95% UI: -0.399 to -0.224) to 0.478 SMD (95% UI: 0.437 to 0.541). In the case of the quadratic component, the 95%HI Full ranges from − 0.021 SMD (95% UI: -0.025 to -0.018) to 0.016 SMD (95% UI: 0.013 to 0.022). Variability within Individuals The median variability within individuals was 0.121 SMD and ranged from 0.066 to 0.277. This variation was positively correlated with the magnitude of the individuals’ average effect (r = 0.39, 95% confidence interval 0.17 to 0.58). Only 13 individuals (19%) showed a negative parameter estimate in more than half of the exercises. Of those patients, one exhibited 8 out of 8 negative parameter estimates, and another individual 7 out of 8 negative parameter estimates. Twenty-one (31%) displayed positive parameter estimates for all exercises and 17 (25%) only one negative parameter estimate. Figure 3 exemplarily displays the individual posterior distribution for 3 individuals. Training sessions until plateau After 20.6 TS (95% UI: 14.8 to 44.4) the average function reaches its vertex. Across exercises the vertex is reached after a posterior median of 13.4 TS (5.7 to 47.6) for butterfly and 22.9 TS (13.1 to 68.9) for knee flexion (Fig. 4). As notable in Table 4 the 95% UI display large overlap, thus, a confident interpretation of seems premature. Discussion Within the last two decades RT has emerged as a valuable support measure for cancer patients in fighting tumor and therapy associated side effects [ 9 ]. However, these benefits are subject to substantial variation. To the best of our knowledge this is the first study that systematically investigates RT response variability in breast cancer patients undergoing cancer therapy. We fitted a three-level hierarchical model to the data and estimated separate regression models for each exercise and individual, besides the average trend. We chose the baseline standardized training volume as dependent variable, which is constituted by the load, the number of sets and the number of repetitions per set, since changes in either one of the constituents is an indicator of training progression. As expected, we observed positive average responses for all exercises. Additionally, the analysis revealed variability between exercises with a coefficient of variation of about 25% with respect to the average effect. This variation is also evident in the different proportion of individuals responding to the particular exercises. For instance the butterfly exercise displayed a positive response in all but one participant (98%) whereas roughly a quarter of participants (77%) showed no increase in the rowing exercise. With regards to the individual responses, only one case showed a negative time trajectory across all exercises. Thirteen more cases (19%) showed a decline in the at least half of the exercises, whereas more than half of participants (55%) displayed no more than one negative trend across exercises. Still, the high proportion of adverse responses appears surprising at first, however, it is crucial to acknowledge this analysis was conducted in cancer patients undergoing adjuvant treatment. Tumors and their treatment inhibit anabolic pathways while enhancing catabolic pathways in the muscle cell [ 23 ]. In a recent systematic review of randomized controlled trials, we investigated the change in body composition in cancer patients undergoing exercise therapy [ 20 ]. In a subgroup-analysis we observed a pooled loss in lean body mass in the non-exercising control group, which is in line with other observations [ 30 ]. This indicates that, in contrast to the general population where one would expect a maintenance in fitness over a relatively short period of time, in cancer patients a decline in muscle mass[ 30 ] and function has to be expected if not actively counteracted via exercise [ 14 , 19 ]. From this perspective, the high number of positive responses indicates that most patients did not only prevent the expected functional and structural decline but overcame the negative trend. Thus, instead of dichotomizing participants in responders and non-responders, we suggest to trichotomize the response continuum in clinical populations that are at risk of losing muscle quality. 1st those who are insensitive to the stimulus and align with what would be expected in the control condition, 2nd those who are successful in preventing the decline in strength and muscle mass, and 3rd those who respond to the stimulus by advancing beyond baseline. Therefore, in a population at risk of accelerated loss of muscle mass and strength, avoiding any decline of the magnitude of the counterfactual control appears to be an accomplishment. As hypothesized, the analysis yielded a positive linear and a negative quadratic component for all exercises, resulting in important practical implications. First, the model revealed that all exercises reach a plateau after about 20 TS or 10 weeks with two weekly TSs. Considering that the majority of studies in exercise oncology and the corresponding guidelines[ 9 , 35 ] neglect to periodize the exercise regimes, as it is recommended in healthy and athletic populations [ 31 ], might explain why several systematic reviews have failed to identify a positive relationship between the duration of a study and the increase of muscle functioning or strength beyond 12 weeks of intervention duration [ 22 , 24 , 38 ]. Second, conventional progression approaches, such as increasing the load by a fixed weight or increasing the load by a fixed proportion of the training load (e.g. 2–10%) [ 31 ] mismatches the empirical volume-time trajectory. This incongruence might lead to a non-optimal load of the muscles. Based on our analysis, the magnitude of the overload should be reduced with each progression, by a small increment of approximately 4% of the initial progression of roughly 0.1 SMD which in our data equals roughly 3–5% of the initial load. Future studies should investigate the efficacy of periodization and progressive overload schemas via confirmatory study designs in cancer patients. There are several limitations to our study. First, it is extremely difficult to operationalize response to resistance exercises in a comprehensive manner [ 15 , 39 , 40 ]. Volume as outcome variable has the advantage that it incorporates several progression indicators. However, all three indicators are not independent and a change in one will automatically be reflected in the others. Second, we chose a quadratic function because of the curvilinear behavior of training adaptations. This allowed us to calculate the vertex of the function and to estimate when the training trajectory reaches its plateau. Nevertheless, the quadratic function, since it drops after the vertex, does not mimic the real training trajectory. More precise trajectories could be modeled by applying a proper growth function. Third, the impact of potential predictor variables such as age, therapy status and severity of the disease need to be investigated in further research to improve the personalization of exercise routines and provide therapists with more realistic expectations about their patients’ improvements. Ultimately, improvements in patient centered variables, such as quality of life or cancer related fatigue, are the primary goal of exercise intervention, whereas increases in training variables are only of subordinate interest. Thus, to optimize training programs for cancer patients we need a better understanding of the causal relationship between resistance training, strength, and other health outcomes. Conclusion Despite some variation, breast cancer patients undergoing cancer treatment can improve their strength without limitations to specific exercises and independent of the recruited muscles. Still, therapists should be aware of the demonstrated differences in patients' training progression and the non-linearity of the training progression so that they act as needed. Eventually, a more personalized approach is needed, which requires closer monitoring of patients and a high expertise in therapists with the fundamentals of RT and the utilization of alternative RT methods [ 5 ]. Declarations Funding Statement: The BEST-Study was funded by the Interdisciplinary Research Funding Program (intramural) of the National Center for Tumor Diseases (NCT), Heidelberg, Germany (grant number IFP project VI.1); and the foundations ‘Stiftung Leben mit Krebs’ and the ‘Manfred-Lautenschlaeger-Stiftung’ that partially supported our intervention programs. The BEATE-Study The study was funded by the German Cancer Research Center (DKFZ), Division of Preventive Oncology. The interventions were partially supported by the foundation “Stiftung Leben mit Krebs”. Conflict of Interest Disclosure : All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript. Author Contribution MK developed the research question, performed the analyses, interpreted the results and drafted the manuscript. JW developed the research question and interpreted the results. MS participated in the design of the study and supported the statistical analysis. FR developed the research question and interpreted the results. KS conceived and supervised the whole study process. All authors helped to draft the manuscript and approved the final manuscript. References AhtiainenJP,WalkerS,PeltonenH,HolvialaJ,SillanpääE,KaravirtaL,SallinenJ,MikkolaJ,ValkeinenH,MeroA(2016)Heterogeneityinresistancetraining-inducedmusclestrengthandmassresponsesinmenandwomenofdifferentagesAge38:1–13 AtkinsonG,WilliamsonP,BatterhamAM(2019)Issues in the determination of ‘responders’ and ‘non-responders’inphysiological research Experimental Physiology104:1215–1225 BarbalhoMdSM,GentilP,IzquierdoM,FisherJ,SteeleJ,deAzevedo RaiolR(2017)Therearenono-responderstoloworhighresistancetrainingvolumesamongolderwomenExperimentalgerontology99:18–26 BarnettAG,Van DerPolsJC,DobsonAJ(2005)Regression to the mean: what it is and how to deal with itInternational journal of epidemiology34:215–220 BettarigaF,BishopC,TaaffeDR,GalvãoDA,MaestroniL,NewtonRU(2023)Time to consider the potential role of alternative resistance training methods in cancer management?Journal of Sport and Health Science BolgerN,ZeeKS,Rossignac-MilonM,HassinRR(2019)Causal processes in psychology are heterogeneousJournal of experimental psychology: General148:601 BrzyckiM(1993)Strength testing—predicting a one-rep max from reps-to-fatigueJournal of physical education, recreation & dance64:88–90 BürknerP-C(2017)brms: An R package for Bayesian multilevel models using StanJournal of statistical software80:1–28 CampbellKL,Winters-StoneK,WiskemannJ,MayAM,SchwartzAL,CourneyaKS,ZuckerD,MatthewsC,LigibelJ,GerberL(2019)Exerciseguidelinesforcancersurvivors:consensusstatementfrominternationalmultidisciplinaryroundtableMedicineandscienceinsportsandexercise51:2375 Churchward-VenneTA,TielandM,VerdijkLB,LeendersM,DirksML,DeGrootLC,VanLoonLJ(2015)There are no nonresponders to resistance-type exercise training in older men and womenJournal of the American Medical Directors Association16:400–411 CourneyaKS,McKenzieDC,MackeyJR,GelmonK,ReidRD,FriedenreichCM,LadhaAB,ProulxC,VallanceJK,LaneK(2008)Moderators of the effects of exercise training in breast cancer patients receiving chemotherapy: a randomized controlled trial Cancer:Interdisciplinary International Journal of the American Cancer Society112:1845–1853 DepaoliS,Van deSchootR(2017)ImprovingtransparencyandreplicationinBayesianstatistics:TheWAMBS-ChecklistPsychologicalmethods22:240 ErskineRM,JonesDA,WilliamsAG,StewartCE,DegensH(2010)Inter-individual variability in the adaptation of human muscle specific tension to progressive resistance training Europeanjournal of applied physiology110:1117–1125 FelicianoEMC,VasanS,LuoJ,BinderAM,ChlebowskiRT,QuesenberryC,BanackHR,CaanBJ,PaskettED,WilliamsGR(2023)Long-termtrajectoriesofphysicalfunctiondeclineinwomenwithandwithoutcancerJAMAoncology9:395–403 GennerKM,WestonM(2014)A comparison of workload quantification methods in relation to physiological responses to resistance exerciseThe Journal of Strength & Conditioning Research28:2621–2627 HeckstedenA,KraushaarJ,Scharhag-RosenbergerF,TheisenD,SennS,MeyerT(2015)Individual response to exercise training-a statistical perspectiveJournal of applied physiology118:1450–1459 HubalMJ,Gordish-DressmanH,ThompsonPD,PriceTB,HoffmanEP,AngelopoulosTJ,GordonPM,MoynaNM,PescatelloLS,VisichPS(2005)Variability in muscle size and strength gain after unilateral resistance trainingMedicine & science in sports & exercise37:964–972 JonesLW(2015)Precision oncology framework for investigation of exercise as treatment for cancerJournal of Clinical Oncology33:4134 KlassenO,SchmidtME,UlrichCM,SchneeweissA,PotthoffK,SteindorfK,WiskemannJ(2017)Muscle strength in breast cancer patients receiving different treatment regimesJournal of cachexia, sarcopenia and muscle8:305–316 KoeppelM,MathisK,SchmitzKH,WiskemannJ(2021)Muscle hypertrophy in cancer patients and survivors via strength training. A meta-analysis and meta-regressionCritical reviews in oncology/hematology163:103371 KravitzRL,DuanN,BraslowJ(2004)Evidence-based medicine, heterogeneity of treatment effects, and the trouble with averagesThe Milbank Quarterly82:661–687 LeeJ(2021)The effects of resistance training on muscular strength and hypertrophy in elderly cancer patients: A systematic review and meta-analysisJournal of Sport and Health Science LondheP,GuttridgeDC(2015)Inflammation induced loss of skeletal muscleBone80:131–142 LopezP,TaaffeDR,NewtonRU,GalvaoDA(2021)Resistanceexercisedosageinmenwithprostatecancer:systematicreview,meta-analysis,andmeta-regressionMedicineandscienceinsportsandexercise53:459 MannTN,LambertsRP,LambertMI(2014)High responders and low responders: factors associated with individual variation in response to standardized trainingSports Medicine44:1113–1124 MunnJ,HerbertRD,GandeviaSC(2004)Contralateral effects of unilateral resistance training: a meta-analysisJournal of applied physiology96:1861–1866 NishinoM(2018)Tumor response assessment for precision cancer therapy: response evaluation criteria in solid tumors and beyond American Society of ClinicalOncology Educational Book38:1019–1029 PearceyG,AlizedahS,PowerK,ButtonD(2021)Chronic resistance training: is it time to rethink the time course of neural contributions to strength gain?European Journal of Applied Physiology121:2413–2422 PotthoffK,SchmidtME,WiskemannJ,HofH,KlassenO,HabermannN,BeckhoveP,DebusJ,UlrichCM,SteindorfK(2013)Randomized controlled trial to evaluate the effects of progressive resistance training compared to progressive muscle relaxation in breast cancer patients undergoing adjuvant radiotherapy: the BEST studyBMC cancer13:1–11 PradoCM,AntounS,SawyerMB,BaracosVE(2011)Two faces of drug therapy in cancer: drug-related lean tissue loss and its adverse consequences to survival and toxicity Current Opinion in Clinical Nutrition &Metabolic Care14:250–254 RatamessNA,AlvarBA,EvetochTE,HoushTJ,Ben KiblerW,KraemerWJ,TriplettNT(2009)Progression models in resistance training for healthy adults Medicine and science in sports and exercise41:687–708 SaleDG(1988)NeuraladaptationtoresistancetrainingMedicine&ScienceinSports&Exercise20 SchmidtME,WiskemannJ,ArmbrustP,SchneeweissA,UlrichCM,SteindorfK(2015)Effects of resistance exercise on fatigue and quality of life in breast cancer patients undergoing adjuvant chemotherapy: a randomized controlled trialInternational journal of cancer137:471–480 SchmidtME,WiskemannJ,Krakowski-RoosenH,KnickerAJ,HabermannN,SchneeweissA,UlrichCM,SteindorfK(2013)Progressive resistance versus relaxation training for breast cancer patients during adjuvant chemotherapy: design and rationale of a randomized controlled trial (BEATE study)Contemporary clinical trials34:117–125 SchmitzKH,CourneyaKS,MatthewsC,Demark-WahnefriedW,GalvãoDA,PintoBM,IrwinML,WolinKY,SegalRJ,LuciaA(2010)American college of sports medicine roundtable on exercise guidelines for cancer survivorsMedicine & Science in Sports & Exercise42:1409–1426 ScottJM,NilsenTS,GuptaD,JonesLW(2018)Exercise therapy and cardiovascular toxicityin cancer Circulation137:1176–1191 SteindorfK,SchmidtM,KlassenO,UlrichC,OelmannJ,HabermannN,BeckhoveP,OwenR,DebusJ,WiskemannJ(2014)Randomized, controlled trial of resistance training in breast cancer patients receiving adjuvant radiotherapy: results on cancer-related fatigue and quality of lifeAnnals of oncology25:2237–2243 SweegersMG,AltenburgTM,BrugJ,MayAM,VanVulpenJK,AaronsonNK,ArbaneG,BohusM,CourneyaKS,DaleyAJ(2019)Effects and moderators of exercise on muscle strength, muscle function and aerobic fitness in patients with cancer: a meta-analysis of individual patient dataBritish journal of sports medicine53:812–812 TanB(1999)Manipulating resistance training program variables to optimize maximum strength in men: a reviewThe Journal of Strength & Conditioning Research13:289–304 ToigoM,BoutellierU(2006)New fundamental resistance exercise determinants of molecular and cellular muscle adaptations Europeanjournal of applied physiology97:643–663 van deSchootR,DepaoliS,KingR,KramerB,MärtensK,TadesseMG,VannucciM,GelmanA,VeenD,WillemsenJ(2021)Bayesian statistics and modellingNature Reviews Methods Primers1:1–26 Tables Tables 1-4 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1.docx Table2.docx Table3.docx Table4.docx SI1OutlierAnalysis.docx SI2StatisticalModel.docx SI3PriorSpecificationandPPC.docx SI4DiffusePrior.docx SI5Comparisonofquadraticandlineartwolevelmodel.docx SI6AdherenceData.docx SI7Frequencyofpositiveposteriormeansperexercise.docx SI8Variationbetweenindividualsforbutterflyandrowingexercise.pptx SI9CERTChecklist.xlsx Cite Share Download PDF Status: Published Journal Publication published 10 Dec, 2024 Read the published version in Supportive Care in Cancer → Version 1 posted Editorial decision: Revision requested 18 Aug, 2024 Reviews received at journal 16 Aug, 2024 Reviews received at journal 08 Aug, 2024 Reviews received at journal 06 Aug, 2024 Reviewers agreed at journal 31 Jul, 2024 Reviewers agreed at journal 26 Jul, 2024 Reviewers agreed at journal 26 Jul, 2024 Reviewers agreed at journal 26 Jul, 2024 Reviewers agreed at journal 13 May, 2024 Reviewers invited by journal 27 Apr, 2024 Editor assigned by journal 21 Apr, 2024 Submission checks completed at journal 15 Mar, 2024 First submitted to journal 13 Mar, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About In Review Editorial Policies Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4089501","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":279843078,"identity":"5f56a424-1c2d-4401-8cbc-3b2db9ca1fc4","order_by":0,"name":"Maximilian Koeppel","email":"","orcid":"","institution":"Heidelberg University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Maximilian","middleName":"","lastName":"Koeppel","suffix":""},{"id":279843080,"identity":"40cb1da0-5c0f-4839-b851-4b61341e75ed","order_by":1,"name":"Karen Steindorf","email":"","orcid":"","institution":"German Cancer Research Center","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Karen","middleName":"","lastName":"Steindorf","suffix":""},{"id":279843082,"identity":"5ef62d90-cbc7-4d8c-b215-880c13f232c6","order_by":2,"name":"Martina Schmidt","email":"","orcid":"","institution":"German Cancer Research Center","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Martina","middleName":"","lastName":"Schmidt","suffix":""},{"id":279843084,"identity":"dcc25c2f-e0df-4fc7-8a2d-171981e30886","order_by":3,"name":"Friederike Rosenberger","email":"","orcid":"","institution":"National Center for Tumor Diseases","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Friederike","middleName":"","lastName":"Rosenberger","suffix":""},{"id":279843087,"identity":"7e63d867-d376-4bfc-940c-33348052ee91","order_by":4,"name":"Joachim Wiskemann","email":"data:image/png;base64,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","orcid":"","institution":"National Center for Tumor Diseases","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Joachim","middleName":"","lastName":"Wiskemann","suffix":""}],"badges":[],"createdAt":"2024-03-13 05:46:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4089501/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4089501/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s00520-024-09001-4","type":"published","date":"2024-12-10T15:57:14+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":53015705,"identity":"3df1ab0f-a914-4b35-9cc0-3d74ff6ba6d7","added_by":"auto","created_at":"2024-03-19 15:58:44","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":118439,"visible":true,"origin":"","legend":"\u003cp\u003ePatient Flow\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;C\u003cem\u003eaption\u003c/em\u003e: \u003csup\u003ea\u003c/sup\u003eHD: Heidelberg, \u003csup\u003eb\u003c/sup\u003e100% refers to the maximum of 24 training sessions)\u003c/p\u003e","description":"","filename":"Slide1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/34c8e336724a5c5a810a85aa.jpg"},{"id":53013961,"identity":"f04ca7f5-ac91-4cfb-a3a6-bd2c103c7794","added_by":"auto","created_at":"2024-03-19 15:50:44","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":121934,"visible":true,"origin":"","legend":"\u003cp\u003ePosterior distribution of a) the linear and b) quadratic component for each exercise.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCaption\u003c/em\u003e: The dashed black line marks the zero line, the dashed red line the population level estimate.\u003c/p\u003e","description":"","filename":"Slide2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/cbba8782754a3c1d894c3945.jpg"},{"id":53013959,"identity":"7af2e1b5-b35e-4db5-ac36-37604ae028f5","added_by":"auto","created_at":"2024-03-19 15:50:44","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":123593,"visible":true,"origin":"","legend":"\u003cp\u003eVariation of posterior distributions per exercise within individuals. Exemplary plots for 3 patients: a) Above average response, b) average response, c) below average response\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCaption\u003c/em\u003e: The dashed black line marks the zero line, the dashed red line the population level estimate. A blue area under the curve characterizes positive parameter values, whereas a grey area under the curve characterizes negative values\u003c/p\u003e","description":"","filename":"Slide3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/7ae1fa4280aae6ac25b667f4.jpg"},{"id":53013963,"identity":"edd5e1e9-b86b-4138-9183-01b8d2f37cca","added_by":"auto","created_at":"2024-03-19 15:50:44","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":133431,"visible":true,"origin":"","legend":"\u003cp\u003ePredicted quadratic trajectory for each exercise\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eCaption: Due to the quadratic model the Training volume will decrease after it reached the peak. This is certainly not a depiction of the real training trajectory.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"Slide4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/9e4ae5e95bdd1a77afbbf21e.jpg"},{"id":71552549,"identity":"d67a93fe-ef47-4178-b685-059c565ffb56","added_by":"auto","created_at":"2024-12-16 16:07:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":929189,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/e7a90b58-3859-4d05-a7ee-f48afd363dd3.pdf"},{"id":53015706,"identity":"fb520fd2-9132-4fbe-9e1a-fc05cc8fb67e","added_by":"auto","created_at":"2024-03-19 15:58:44","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":10487,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/d126f7b8369ff6f267c27a41.docx"},{"id":53013965,"identity":"4e7d2a92-e96c-43ee-b9ff-36f7f6fb9d20","added_by":"auto","created_at":"2024-03-19 15:50:44","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":18521,"visible":true,"origin":"","legend":"","description":"","filename":"Table2.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/928b23e6255241af20891987.docx"},{"id":53015707,"identity":"1cf4d7d2-89cb-4bc4-8661-8af157415449","added_by":"auto","created_at":"2024-03-19 15:58:44","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":24655,"visible":true,"origin":"","legend":"","description":"","filename":"Table3.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/923b38ada21cae2affaae710.docx"},{"id":53015708,"identity":"45480403-fa1f-4908-9acc-d4f3d6ec74da","added_by":"auto","created_at":"2024-03-19 15:58:44","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":18701,"visible":true,"origin":"","legend":"","description":"","filename":"Table4.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/940cb90e18bb182faadb9821.docx"},{"id":53013966,"identity":"2a55f96e-99a8-486f-a9b3-b706165a69df","added_by":"auto","created_at":"2024-03-19 15:50:44","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":29836,"visible":true,"origin":"","legend":"","description":"","filename":"SI1OutlierAnalysis.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/5d35bbccd3fe15b69077d08a.docx"},{"id":53013968,"identity":"57cacf14-8afc-4c57-a1c8-a9dfcf0f2cd6","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":19343,"visible":true,"origin":"","legend":"","description":"","filename":"SI2StatisticalModel.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/24c471de76507574a0206eee.docx"},{"id":53015709,"identity":"5cae6147-4cae-498d-bde1-ffb7b6ac320b","added_by":"auto","created_at":"2024-03-19 15:58:45","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":815846,"visible":true,"origin":"","legend":"","description":"","filename":"SI3PriorSpecificationandPPC.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/663f1e5f2657df70251d491b.docx"},{"id":53013969,"identity":"51d2a0cc-dea0-47cd-be26-7b8bd8f7c884","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"docx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":21650,"visible":true,"origin":"","legend":"","description":"","filename":"SI4DiffusePrior.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/a06e563f1af1988a8af26c96.docx"},{"id":53013971,"identity":"6ea19cd7-3865-4ba5-9771-72b88c0f6aff","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":28554,"visible":true,"origin":"","legend":"","description":"","filename":"SI5Comparisonofquadraticandlineartwolevelmodel.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/af605b60a5d45b56786a3f54.docx"},{"id":53015710,"identity":"0a1679f5-ad5d-4097-a41c-dddedb0d4733","added_by":"auto","created_at":"2024-03-19 15:58:45","extension":"docx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":26287,"visible":true,"origin":"","legend":"","description":"","filename":"SI6AdherenceData.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/1c9b24b7d2c356a61ff963fe.docx"},{"id":53013975,"identity":"2cee8f2a-7e27-49fc-8d00-41b072b4737b","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"docx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":19040,"visible":true,"origin":"","legend":"","description":"","filename":"SI7Frequencyofpositiveposteriormeansperexercise.docx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/59a1711f292ec787bece3756.docx"},{"id":53013974,"identity":"5ce2d5bb-ce49-487e-9809-f81408836d3f","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"pptx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":3127963,"visible":true,"origin":"","legend":"","description":"","filename":"SI8Variationbetweenindividualsforbutterflyandrowingexercise.pptx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/6a8ccf13afa2a9599bed1b33.pptx"},{"id":53013973,"identity":"8fe16e14-14a4-422c-be03-9bf10e68ada0","added_by":"auto","created_at":"2024-03-19 15:50:45","extension":"xlsx","order_by":15,"title":"","display":"","copyAsset":false,"role":"supplement","size":9158,"visible":true,"origin":"","legend":"","description":"","filename":"SI9CERTChecklist.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-4089501/v1/81bbffc3e62f9a33a88a6ad3.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Variability in resistance training trajectories of breast cancer patients undergoing therapy","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBased on a large body of evidence, resistance training provides several positive effects for cancer patients. However, RT effects do not follow a simple causal stimulus-response relationship between mechanical input and physiological adaptation, but are subject to a complex network of effect modificators [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Thus, RT effects may vary between individuals and leads to the classification of individuals into distinct response-categories [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. In contrast to medical oncology, in which response refers to the efficacy of tumor treatment in reducing tumor size or severity [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] the in RT research the term is used more ambiguously and could refer to several RT related outcomes of interest, such as the one repetition maximum (1RM), the cross sectional area of a particular muscle [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], or the performance in a functional test [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis observations has stimulated scientific endeavors over the last few years and gave rise to controversial discussions. One major criticism is that studies are often confined to comparing data of only two distinct time points with each other. Where the first point is timed before the start of the exercise intervention and the second one right after the end of the training period. Thus, the two measurements are prone to within-individual variation, such as different mood, motivation and pain tolerance, but also measurement error which can lead to misestimations [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Another statistical artifact, would be that extreme values in the first assessment tend to be drawn closer to the mean in the second assessment, resulting in an overestimation of the effect variability [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. From a more practical perspective, it is well known that strength gains follow a non-linear time-trajectory, with the steepest incline at the beginning of the intervention, which eventually approaches a plateau during the course of the intervention [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. By ignoring the qualitative differences, linear models tend to underestimate the strength gains in the early stages of training and overestimate the later ones. Furthermore, strength gains do not appear systemic but locally, involving primarily those muscles engaged in the particular exercise [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Therefore, it is surprising that the RT response variability within individuals has not received much attention so far.\u003c/p\u003e \u003cp\u003eAlthough response variability to exercise has been mentioned in several well placed publications, in exercise oncology [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] it did not attract the researchers\u0026rsquo; focus of interest. However, the investigation of exercise response variability is especially important for clinical populations, to ensure the optimal care of these populations. Ignoring the effect variability could lead to false expectations for the individual patient, due to an overconfidence in a one-size-fits-all approach [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIt is the purpose of this study to investigate how the change in training volume varies between exercises and between individuals undergoing adjuvant cancer treatment over the course of an RT intervention. In addition, we are interested if, and at what point in time these training-trajectories might reach a plateau.\u003c/p\u003e \u003cp\u003eDue to the nested structure of the data we fitted a hierarchical Bayesian regression model, which allowed us to estimate the variability between exercises and between individuals. In order to model a potentially non-linear time-volume relationship and estimate the number of TS leading to a plateau, we included a quadratic term to the model.\u003c/p\u003e"},{"header":"Method","content":"\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eDesign and participants\u003c/h2\u003e \u003cp\u003eWe conducted a secondary analysis by pooling the data of the BEATE (NCT01106820) and the BEST trial (NCT01468766), two randomized clinical trials conducted with physically inactive (\u0026lt;\u0026thinsp;1h/week exercising), non-metastatic breast cancer patients (age\u0026thinsp;\u0026gt;\u0026thinsp;18 years) undergoing adjuvant chemotherapy (BEATE trial) or adjuvant radiation therapy (BEST trial). Detailed information about the study design are described elsewhere [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Briefly summarized, both trials investigated the effect of a 12-week machine-based RT intervention on cancer-related fatigue. These studies were approved by the ethics committee (S-012/2009, S-447/2010), were conducted in accordance with the Declaration of Helsinki, and written informed consent was obtained from all patients.\u003c/p\u003e \u003cp\u003eA total of 261 patients (101 BEATE, 160 BEST) were enrolled in both trials with 132 assigned to the intervention group (52 BEATE, 80 BEST). While many patients trained at the university\u0026rsquo;s training facility (41 BEATE, 31 BEST), the patients were also given the option to train in facilities near their homes in order to reduce the barriers for participation. For this analysis, we only included the 69 patients who trained at the study center\u0026rsquo;s facility to reduce additional sources of variance (Fig.\u0026nbsp;1). Baseline patient characteristics are outlined in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eIntervention\u003c/h2\u003e \u003cp\u003eIn accordance with the ACSM exercise guidelines for cancer survivors, current during the conduction of the study [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e], the training was planned to take place twice a week over the course of 12 weeks, resulting in a maximum number of 24 TS. In both studies the same core exercises were conducted: leg press, leg extension, leg curl, shoulder internal and external rotation, seated row, latissimus pull down and butterfly. Patients who participated in the BEATE trial also performed butterfly reverse each TS and anteversion or retroversion of the shoulder every other TS. Each exercise was conducted for 1 to 3 sets with 12 repetitions at 60\u0026ndash;80% of their hypothetical one repetition maximum, estimated with the Brzycki-Formula [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. The training schedule followed a progressive approach, in which the applied load was increased by at least 5% if the prescribed load was successfully lifted for 3 sets of 12 repetitions in three consecutive TS. The TS were supervised by experienced exercise- or physical therapists. For more details please refer to the design [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] and primary endpoint papers [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e] as well as the CERT-Checklist (Supplementary Information SI 9).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of outcome and predictor variables\u003c/h2\u003e \u003cp\u003eThe goal of the analysis was to investigate the change in training volume over time. Therefore the number of the TS was chosen as predictor variable. We excluded the first two TSs from each patient since those were considered familiarization sessions. Training volume per TS (product of load, number of sets and number of repetitions realized in the particular session) was chosen as our best available proxy in determining the training progression. The training volume per TS for each exercise was z-standardized to the baseline mean and baseline standard deviation of the particular exercise and served as outcome variable for all further analyses. Thus, the resulting regression coefficients can be interpreted and will be referred to as standardized mean differences (SMD). TS were considered as valid if more than eight repetitions and more than two sets were performed. If the number of repetitions exceeded 12, training volume was equalized with the volume of the next valid session to avoid misestimations. Management of outliers is discussed in (Supplementary Information SI 1).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eWe applied a quadratic hierarchical Bayesian regression model with training volume as outcome variable and number of TS as predictor variable. We added a quadratic term to the model, assuming that training progression does not follow a linear path but will display a steep incline in the first training sessions and eventually approaching a plateau [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. The model consists of three levels: The individual TS (Level 1) which are assigned to the different exercises (Level 2), which are then assigned to the patients (Level 3). This model allowed all parameters (i.e. intercept, linear component and quadratic component) to vary on the 2nd and 3rd level and enabled us to investigate the variability of the model parameters between exercises (Level 2), between individuals (Level 3) but also the between exercises within individuals. Regarding the variability of the RT response trajectories, we focus on the linear component, i.e. a positive linear component describes a positive response, and a negative linear component describes an adverse response.\u003c/p\u003e \u003cp\u003eWe chose a Bayesian framework since it allowed us to integrate a large number of already existing information into the prior distribution. Precisely we used the data of an unpublished meta-analysis of 25 resistance training studies with breast cancer patients and survivors, including a total of 112 effect sizes (Supplementary Information SI 3). Furthermore the Bayesian analysis does not result in a point estimate for each parameter but in a probability distribution, which allowed us probabilistic interpretations of the parameter estimates[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. To summarize the posterior distribution of a parameter, we chose the posterior distribution\u0026rsquo;s mean as the most probable location of the parameter and its 95% uncertainty interval (95% UI), thus, describing the range of the parameter values where the true value can be expected with a 95% probability. For simplicity\u0026rsquo;s sake, we will be referring to the mean of the posterior distribution as the parameter estimate if not otherwise specified. The analysis was conducted via in R with the brms package[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e95% Heterogeneity Intervals (95% HI)\u003c/h2\u003e \u003cp\u003eIn accordance with Bolger et al. [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e] we also calculated the 95% HI for the exercises (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({95\\%HI}_{Exercise}\\)\u003c/span\u003e\u003c/span\u003e) and the full model (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({95\\%HI}_{Full}\\)\u003c/span\u003e\u003c/span\u003e). The \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({95\\%HI}_{Exercise}\\)\u003c/span\u003e\u003c/span\u003e states (under the assumption of normality) that the parameter estimate of any exercise for this specific population is expected to lie within this interval with a probability of 95%. For clarification, this differs from the definition of the 95% UI, since the 95% UI describes where the parameter estimate of one exercise can be expected with a 95% probability. Analogously, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({95\\%HI}_{Full}\\)\u003c/span\u003e\u003c/span\u003e, refers to the range of values where any parameter estimate, regardless of person or exercise, would be expected with a probability of 95%.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eTraining Sessions until Plateau (TS\u003csub\u003ePlateau\u003c/sub\u003e)\u003c/h2\u003e \u003cp\u003eThe first derivative of the quadratic regression formula allows to calculate the vertex of the function:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$${TS}_{Plateau}=\\frac{LinearComponent}{2*QuadraticComponent}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThis vertex serves as the indicator, where the trajectory will reach its plateau. To model this vertex, 10,000 values were randomly drawn from the posterior distributions of the linear and the quadratic component, resulting in 10,000 pairs of estimates. Each of these pairs was entered into the formula introduced above. This in turn resulted in a probability distribution consisting of 10,000 values for \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\({TS}_{Plateau}\\)\u003c/span\u003e\u003c/span\u003e. In case of profoundly skewed posterior distributions, we chose to report the median of the posterior distribution. To ensure the transparency and reliability of the analysis, we followed the WAMBS-Checklist [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] (Supplementary Information SI 4 for the analysis with diffuse priors). In addition, we fitted linear two-level-models to the data of each exercise and compared them to their quadratic counterparts, using leave-one-out cross validation (Supplementary Information SI 5) to justify the utilization of the quadratic model. Finally, we reran the three-level-model with only those individuals attending at least 50% of TSs (n\u0026thinsp;=\u0026thinsp;38, 53%) and the ones attending at least 75% (n\u0026thinsp;=\u0026thinsp;27, 38%) of TS (Fig.\u0026nbsp;1) to check if the parameter estimates are skewed by systematically low attendance (Supplementary Information SI 6).\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eThe raw baseline values for all exercises are provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eSummary statistics, such as the means, standard deviations (SD) and 95% UI for the posterior distributions from the main analysis are displayed in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. A comparison between the linear and the quadratic model showed a superiority of the quadratic model (Supplementary Information SI 5), hence, only the results from the quadratic model will be displayed in the main manuscript.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMain Analysis\u003c/h2\u003e \u003cp\u003eThe main analysis yielded an average linear component of SMD\u0026thinsp;=\u0026thinsp;0.093 (95%UI: 0.058 to 0.120) per TS and an average quadratic component of SMD= -0.002 (95%UI: -0.008 to -0.001) per TS-squared.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eVariability between Exercises\u003c/h2\u003e \u003cp\u003eBetween exercises we observed a standard deviation of SD\u0026thinsp;=\u0026thinsp;0.043 (95%UI: 0.018 to 0.082) for the linear component and SD\u0026thinsp;=\u0026thinsp;0.002 (95%UI: \u0026lt;0.001 to 0.004) for the quadratic component. Figure\u0026nbsp;2 and Table\u0026nbsp;3 show that the posterior distributions of the linear components for all exercises are right of the zero-line, indicating positive training response across exercises. This is also supported by the 95%HI\u003csub\u003eExercise\u003c/sub\u003e, ranging from 0.013 SMD (95%UI: -0.094 to 0.077) to 0.175 SMD (95%UI: 0.140 to 0.234) and a median coefficient of variation at about 45% (95% UI\u0026thinsp;=\u0026thinsp;19 to 88%). The 95%HI\u003csub\u003eExercise\u003c/sub\u003e of the quadratic component ranges from \u0026minus;\u0026thinsp;0.006 SMD (95% UI = -0.009 to -0.003) to 0.001 SMD (95% UI = -0.002 to 0.007). This means that almost any combination of both components results in an inversely L-shaped curve. The variation between exercises also has an impact on the proportion of individuals who display a positive response to a particular exercise, i.e. the proportion of individuals whose posterior distribution\u0026rsquo;s median is positive, ranging from 77% for the rowing exercise to 98% for the butterfly exercise (Supplementary Information SI 7 \u0026amp; 8).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eVariability between Individuals\u003c/h2\u003e \u003cp\u003eBetween individuals the linear component dispersed with a standard deviation of SD\u0026thinsp;=\u0026thinsp;0.155 (95% UI: 0.142 to 0.168) and the quadratic component with SD\u0026thinsp;=\u0026thinsp;0.008 (95% UI: 0.007 to 0.009).\u003c/p\u003e \u003cp\u003eThe variability of the linear component between individuals appears to be 3.6 times the size of the variation between exercises (95% UI: 3.3 to 3.9). Concerning the quadratic component, the inter-individual variability appears to be four times larger than the variation observed in the quadratic components for the different exercises (95% UI: 3.5 to 4.5). Considering the total variation, i.e. individual variation across exercises, the 95%HI\u003csub\u003eFull\u003c/sub\u003e for the linear component ranges from \u0026minus;\u0026thinsp;0.290 SMD (95% UI: -0.399 to -0.224) to 0.478 SMD (95% UI: 0.437 to 0.541). In the case of the quadratic component, the 95%HI\u003csub\u003eFull\u003c/sub\u003e ranges from \u0026minus;\u0026thinsp;0.021 SMD (95% UI: -0.025 to -0.018) to 0.016 SMD (95% UI: 0.013 to 0.022).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eVariability within Individuals\u003c/h2\u003e \u003cp\u003eThe median variability within individuals was 0.121 SMD and ranged from 0.066 to 0.277. This variation was positively correlated with the magnitude of the individuals\u0026rsquo; average effect (r\u0026thinsp;=\u0026thinsp;0.39, 95% confidence interval 0.17 to 0.58). Only 13 individuals (19%) showed a negative parameter estimate in more than half of the exercises. Of those patients, one exhibited 8 out of 8 negative parameter estimates, and another individual 7 out of 8 negative parameter estimates. Twenty-one (31%) displayed positive parameter estimates for all exercises and 17 (25%) only one negative parameter estimate. Figure\u0026nbsp;3 exemplarily displays the individual posterior distribution for 3 individuals.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eTraining sessions until plateau\u003c/h2\u003e \u003cp\u003eAfter 20.6 TS (95% UI: 14.8 to 44.4) the average function reaches its vertex. Across exercises the vertex is reached after a posterior median of 13.4 TS (5.7 to 47.6) for butterfly and 22.9 TS (13.1 to 68.9) for knee flexion (Fig.\u0026nbsp;4). As notable in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e the 95% UI display large overlap, thus, a confident interpretation of seems premature.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eWithin the last two decades RT has emerged as a valuable support measure for cancer patients in fighting tumor and therapy associated side effects [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. However, these benefits are subject to substantial variation. To the best of our knowledge this is the first study that systematically investigates RT response variability in breast cancer patients undergoing cancer therapy. We fitted a three-level hierarchical model to the data and estimated separate regression models for each exercise and individual, besides the average trend. We chose the baseline standardized training volume as dependent variable, which is constituted by the load, the number of sets and the number of repetitions per set, since changes in either one of the constituents is an indicator of training progression. As expected, we observed positive average responses for all exercises. Additionally, the analysis revealed variability between exercises with a coefficient of variation of about 25% with respect to the average effect. This variation is also evident in the different proportion of individuals responding to the particular exercises. For instance the butterfly exercise displayed a positive response in all but one participant (98%) whereas roughly a quarter of participants (77%) showed no increase in the rowing exercise. With regards to the individual responses, only one case showed a negative time trajectory across all exercises. Thirteen more cases (19%) showed a decline in the at least half of the exercises, whereas more than half of participants (55%) displayed no more than one negative trend across exercises. Still, the high proportion of adverse responses appears surprising at first, however, it is crucial to acknowledge this analysis was conducted in cancer patients undergoing adjuvant treatment. Tumors and their treatment inhibit anabolic pathways while enhancing catabolic pathways in the muscle cell [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. In a recent systematic review of randomized controlled trials, we investigated the change in body composition in cancer patients undergoing exercise therapy [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. In a subgroup-analysis we observed a pooled loss in lean body mass in the non-exercising control group, which is in line with other observations [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. This indicates that, in contrast to the general population where one would expect a maintenance in fitness over a relatively short period of time, in cancer patients a decline in muscle mass[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] and function has to be expected if not actively counteracted via exercise [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. From this perspective, the high number of positive responses indicates that most patients did not only prevent the expected functional and structural decline but overcame the negative trend. Thus, instead of dichotomizing participants in responders and non-responders, we suggest to trichotomize the response continuum in clinical populations that are at risk of losing muscle quality. 1st those who are insensitive to the stimulus and align with what would be expected in the control condition, 2nd those who are successful in preventing the decline in strength and muscle mass, and 3rd those who respond to the stimulus by advancing beyond baseline. Therefore, in a population at risk of accelerated loss of muscle mass and strength, avoiding any decline of the magnitude of the counterfactual control appears to be an accomplishment.\u003c/p\u003e \u003cp\u003eAs hypothesized, the analysis yielded a positive linear and a negative quadratic component for all exercises, resulting in important practical implications. First, the model revealed that all exercises reach a plateau after about 20 TS or 10 weeks with two weekly TSs. Considering that the majority of studies in exercise oncology and the corresponding guidelines[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e] neglect to periodize the exercise regimes, as it is recommended in healthy and athletic populations [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], might explain why several systematic reviews have failed to identify a positive relationship between the duration of a study and the increase of muscle functioning or strength beyond 12 weeks of intervention duration [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eSecond, conventional progression approaches, such as increasing the load by a fixed weight or increasing the load by a fixed proportion of the training load (e.g. 2\u0026ndash;10%) [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] mismatches the empirical volume-time trajectory. This incongruence might lead to a non-optimal load of the muscles. Based on our analysis, the magnitude of the overload should be reduced with each progression, by a small increment of approximately 4% of the initial progression of roughly 0.1 SMD which in our data equals roughly 3\u0026ndash;5% of the initial load. Future studies should investigate the efficacy of periodization and progressive overload schemas via confirmatory study designs in cancer patients.\u003c/p\u003e \u003cp\u003eThere are several limitations to our study. First, it is extremely difficult to operationalize response to resistance exercises in a comprehensive manner [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. Volume as outcome variable has the advantage that it incorporates several progression indicators. However, all three indicators are not independent and a change in one will automatically be reflected in the others. Second, we chose a quadratic function because of the curvilinear behavior of training adaptations. This allowed us to calculate the vertex of the function and to estimate when the training trajectory reaches its plateau. Nevertheless, the quadratic function, since it drops after the vertex, does not mimic the real training trajectory. More precise trajectories could be modeled by applying a proper growth function. Third, the impact of potential predictor variables such as age, therapy status and severity of the disease need to be investigated in further research to improve the personalization of exercise routines and provide therapists with more realistic expectations about their patients\u0026rsquo; improvements.\u003c/p\u003e \u003cp\u003eUltimately, improvements in patient centered variables, such as quality of life or cancer related fatigue, are the primary goal of exercise intervention, whereas increases in training variables are only of subordinate interest. Thus, to optimize training programs for cancer patients we need a better understanding of the causal relationship between resistance training, strength, and other health outcomes.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eDespite some variation, breast cancer patients undergoing cancer treatment can improve their strength without limitations to specific exercises and independent of the recruited muscles. Still, therapists should be aware of the demonstrated differences in patients' training progression and the non-linearity of the training progression so that they act as needed. Eventually, a more personalized approach is needed, which requires closer monitoring of patients and a high expertise in therapists with the fundamentals of RT and the utilization of alternative RT methods [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e "},{"header":"Declarations","content":"\u003ch2\u003eFunding Statement:\u003c/h2\u003e \u003cp\u003eThe BEST-Study was funded by the Interdisciplinary Research Funding Program (intramural) of the National Center for Tumor Diseases (NCT), Heidelberg, Germany (grant number IFP project VI.1); and the foundations \u0026lsquo;Stiftung Leben mit Krebs\u0026rsquo; and the \u0026lsquo;Manfred-Lautenschlaeger-Stiftung\u0026rsquo; that partially supported our intervention programs. The BEATE-Study The study was funded by the German Cancer Research Center (DKFZ), Division of Preventive Oncology. The interventions were partially supported by the foundation \u0026ldquo;Stiftung Leben mit Krebs\u0026rdquo;.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003e \u003cb\u003eConflict of Interest Disclosure\u003c/b\u003e:\u003c/strong\u003e \u003cp\u003eAll authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eMK developed the research question, performed the analyses, interpreted the results and drafted the manuscript. JW developed the research question and interpreted the results. MS participated in the design of the study and supported the statistical analysis. FR developed the research question and interpreted the results. KS conceived and supervised the whole study process. All authors helped to draft the manuscript and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAhtiainenJP,WalkerS,PeltonenH,HolvialaJ,Sillanp\u0026auml;\u0026auml;E,KaravirtaL,SallinenJ,MikkolaJ,ValkeinenH,MeroA(2016)Heterogeneityinresistancetraining-inducedmusclestrengthandmassresponsesinmenandwomenofdifferentagesAge38:1\u0026ndash;13\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAtkinsonG,WilliamsonP,BatterhamAM(2019)Issues in the determination of \u0026lsquo;responders\u0026rsquo; and \u0026lsquo;non-responders\u0026rsquo;inphysiological research Experimental Physiology104:1215\u0026ndash;1225\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarbalhoMdSM,GentilP,IzquierdoM,FisherJ,SteeleJ,deAzevedo RaiolR(2017)Therearenono-responderstoloworhighresistancetrainingvolumesamongolderwomenExperimentalgerontology99:18\u0026ndash;26\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarnettAG,Van DerPolsJC,DobsonAJ(2005)Regression to the mean: what it is and how to deal with itInternational journal of epidemiology34:215\u0026ndash;220\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBettarigaF,BishopC,TaaffeDR,Galv\u0026atilde;oDA,MaestroniL,NewtonRU(2023)Time to consider the potential role of alternative resistance training methods in cancer management?Journal of Sport and Health Science\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBolgerN,ZeeKS,Rossignac-MilonM,HassinRR(2019)Causal processes in psychology are heterogeneousJournal of experimental psychology: General148:601\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrzyckiM(1993)Strength testing\u0026mdash;predicting a one-rep max from reps-to-fatigueJournal of physical education, recreation \u0026amp; dance64:88\u0026ndash;90\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB\u0026uuml;rknerP-C(2017)brms: An R package for Bayesian multilevel models using StanJournal of statistical software80:1\u0026ndash;28\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampbellKL,Winters-StoneK,WiskemannJ,MayAM,SchwartzAL,CourneyaKS,ZuckerD,MatthewsC,LigibelJ,GerberL(2019)Exerciseguidelinesforcancersurvivors:consensusstatementfrominternationalmultidisciplinaryroundtableMedicineandscienceinsportsandexercise51:2375\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChurchward-VenneTA,TielandM,VerdijkLB,LeendersM,DirksML,DeGrootLC,VanLoonLJ(2015)There are no nonresponders to resistance-type exercise training in older men and womenJournal of the American Medical Directors Association16:400\u0026ndash;411\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCourneyaKS,McKenzieDC,MackeyJR,GelmonK,ReidRD,FriedenreichCM,LadhaAB,ProulxC,VallanceJK,LaneK(2008)Moderators of the effects of exercise training in breast cancer patients receiving chemotherapy: a randomized controlled trial Cancer:Interdisciplinary International Journal of the American Cancer Society112:1845\u0026ndash;1853\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDepaoliS,Van deSchootR(2017)ImprovingtransparencyandreplicationinBayesianstatistics:TheWAMBS-ChecklistPsychologicalmethods22:240\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErskineRM,JonesDA,WilliamsAG,StewartCE,DegensH(2010)Inter-individual variability in the adaptation of human muscle specific tension to progressive resistance training Europeanjournal of applied physiology110:1117\u0026ndash;1125\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFelicianoEMC,VasanS,LuoJ,BinderAM,ChlebowskiRT,QuesenberryC,BanackHR,CaanBJ,PaskettED,WilliamsGR(2023)Long-termtrajectoriesofphysicalfunctiondeclineinwomenwithandwithoutcancerJAMAoncology9:395\u0026ndash;403\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGennerKM,WestonM(2014)A comparison of workload quantification methods in relation to physiological responses to resistance exerciseThe Journal of Strength \u0026amp; Conditioning Research28:2621\u0026ndash;2627\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeckstedenA,KraushaarJ,Scharhag-RosenbergerF,TheisenD,SennS,MeyerT(2015)Individual response to exercise training-a statistical perspectiveJournal of applied physiology118:1450\u0026ndash;1459\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHubalMJ,Gordish-DressmanH,ThompsonPD,PriceTB,HoffmanEP,AngelopoulosTJ,GordonPM,MoynaNM,PescatelloLS,VisichPS(2005)Variability in muscle size and strength gain after unilateral resistance trainingMedicine \u0026amp; science in sports \u0026amp; exercise37:964\u0026ndash;972\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJonesLW(2015)Precision oncology framework for investigation of exercise as treatment for cancerJournal of Clinical Oncology33:4134\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKlassenO,SchmidtME,UlrichCM,SchneeweissA,PotthoffK,SteindorfK,WiskemannJ(2017)Muscle strength in breast cancer patients receiving different treatment regimesJournal of cachexia, sarcopenia and muscle8:305\u0026ndash;316\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoeppelM,MathisK,SchmitzKH,WiskemannJ(2021)Muscle hypertrophy in cancer patients and survivors via strength training. A meta-analysis and meta-regressionCritical reviews in oncology/hematology163:103371\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKravitzRL,DuanN,BraslowJ(2004)Evidence-based medicine, heterogeneity of treatment effects, and the trouble with averagesThe Milbank Quarterly82:661\u0026ndash;687\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeeJ(2021)The effects of resistance training on muscular strength and hypertrophy in elderly cancer patients: A systematic review and meta-analysisJournal of Sport and Health Science\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLondheP,GuttridgeDC(2015)Inflammation induced loss of skeletal muscleBone80:131\u0026ndash;142\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLopezP,TaaffeDR,NewtonRU,GalvaoDA(2021)Resistanceexercisedosageinmenwithprostatecancer:systematicreview,meta-analysis,andmeta-regressionMedicineandscienceinsportsandexercise53:459\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMannTN,LambertsRP,LambertMI(2014)High responders and low responders: factors associated with individual variation in response to standardized trainingSports Medicine44:1113\u0026ndash;1124\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMunnJ,HerbertRD,GandeviaSC(2004)Contralateral effects of unilateral resistance training: a meta-analysisJournal of applied physiology96:1861\u0026ndash;1866\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNishinoM(2018)Tumor response assessment for precision cancer therapy: response evaluation criteria in solid tumors and beyond American Society of ClinicalOncology Educational Book38:1019\u0026ndash;1029\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePearceyG,AlizedahS,PowerK,ButtonD(2021)Chronic resistance training: is it time to rethink the time course of neural contributions to strength gain?European Journal of Applied Physiology121:2413\u0026ndash;2422\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePotthoffK,SchmidtME,WiskemannJ,HofH,KlassenO,HabermannN,BeckhoveP,DebusJ,UlrichCM,SteindorfK(2013)Randomized controlled trial to evaluate the effects of progressive resistance training compared to progressive muscle relaxation in breast cancer patients undergoing adjuvant radiotherapy: the BEST studyBMC cancer13:1\u0026ndash;11\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePradoCM,AntounS,SawyerMB,BaracosVE(2011)Two faces of drug therapy in cancer: drug-related lean tissue loss and its adverse consequences to survival and toxicity Current Opinion in Clinical Nutrition \u0026amp;Metabolic Care14:250\u0026ndash;254\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRatamessNA,AlvarBA,EvetochTE,HoushTJ,Ben KiblerW,KraemerWJ,TriplettNT(2009)Progression models in resistance training for healthy adults Medicine and science in sports and exercise41:687\u0026ndash;708\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaleDG(1988)NeuraladaptationtoresistancetrainingMedicine\u0026amp;ScienceinSports\u0026amp;Exercise20\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidtME,WiskemannJ,ArmbrustP,SchneeweissA,UlrichCM,SteindorfK(2015)Effects of resistance exercise on fatigue and quality of life in breast cancer patients undergoing adjuvant chemotherapy: a randomized controlled trialInternational journal of cancer137:471\u0026ndash;480\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmidtME,WiskemannJ,Krakowski-RoosenH,KnickerAJ,HabermannN,SchneeweissA,UlrichCM,SteindorfK(2013)Progressive resistance versus relaxation training for breast cancer patients during adjuvant chemotherapy: design and rationale of a randomized controlled trial (BEATE study)Contemporary clinical trials34:117\u0026ndash;125\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchmitzKH,CourneyaKS,MatthewsC,Demark-WahnefriedW,Galv\u0026atilde;oDA,PintoBM,IrwinML,WolinKY,SegalRJ,LuciaA(2010)American college of sports medicine roundtable on exercise guidelines for cancer survivorsMedicine \u0026amp; Science in Sports \u0026amp; Exercise42:1409\u0026ndash;1426\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScottJM,NilsenTS,GuptaD,JonesLW(2018)Exercise therapy and cardiovascular toxicityin cancer Circulation137:1176\u0026ndash;1191\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteindorfK,SchmidtM,KlassenO,UlrichC,OelmannJ,HabermannN,BeckhoveP,OwenR,DebusJ,WiskemannJ(2014)Randomized, controlled trial of resistance training in breast cancer patients receiving adjuvant radiotherapy: results on cancer-related fatigue and quality of lifeAnnals of oncology25:2237\u0026ndash;2243\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSweegersMG,AltenburgTM,BrugJ,MayAM,VanVulpenJK,AaronsonNK,ArbaneG,BohusM,CourneyaKS,DaleyAJ(2019)Effects and moderators of exercise on muscle strength, muscle function and aerobic fitness in patients with cancer: a meta-analysis of individual patient dataBritish journal of sports medicine53:812\u0026ndash;812\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTanB(1999)Manipulating resistance training program variables to optimize maximum strength in men: a reviewThe Journal of Strength \u0026amp; Conditioning Research13:289\u0026ndash;304\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eToigoM,BoutellierU(2006)New fundamental resistance exercise determinants of molecular and cellular muscle adaptations Europeanjournal of applied physiology97:643\u0026ndash;663\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan deSchootR,DepaoliS,KingR,KramerB,M\u0026auml;rtensK,TadesseMG,VannucciM,GelmanA,VeenD,WillemsenJ(2021)Bayesian statistics and modellingNature Reviews Methods Primers1:1\u0026ndash;26\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1-4 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"supportive-care-in-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jscc","sideBox":"Learn more about [Supportive Care in Cancer](https://www.springer.com/journal/520)","snPcode":"520","submissionUrl":"https://submission.nature.com/new-submission/520/3","title":"Supportive Care in Cancer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"exercise oncology, adjuvant tumor treatment, response variability, hierarchical model, Bayesian statistics","lastPublishedDoi":"10.21203/rs.3.rs-4089501/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4089501/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose\u003c/h2\u003e \u003cp\u003eIn resistance training (RT) the change in training volume from training sessions (TS) to TS, is an indicator of training progress. Resulting growth-trajectories are likely to differ between individuals. Understanding this variation is important for exercise planning in general, but even more for clinical populations. We investigated this variation in breast cancer patients undergoing treatment.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eData of 69 patients from two randomized controlled trails were investigated. They conducted a 12-week RT program. We fitted a quadratic Bayesian regression model to the baseline standardized training volume over the course of the intervention. We allowed all parameters to vary both between exercises and between individuals.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eWe observed a positive linear component of 0.093 (95% Uncertainty interval (UI) 0.058 to 0.120) and a negative quadratic component of -0.002 (95% UI -0.008 to 0.001) for the mean trajectory of the change in training volume. For the different exercises we observed a dispersion for both the linear (0.043, 95% UI 0.018 to 0.082) and the quadratic component (0.002, 95% UI\u0026thinsp;\u0026lt;\u0026thinsp;0.001 to 0.004). Variation between-individual appears to be approximately 4 times larger. We also observed between-exercise variation within individuals. Extrapolation of the regression model indicates training progression stagnates after 20.6 TS (95% UI 14.8 to 44.4).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThere is substantial variation in RT response between breast cancer patients undergoing tumor therapy and in-between exercises. The non-linear trajectory indicates that training progression will eventually plateau, demanding periodization and timely modification.\u003c/p\u003e\u003ch2\u003eTrial Registration:\u003c/h2\u003e \u003cp\u003eBEATE Study: NCT01106820, Date: April 20, 2010; BEST Study: NCT01468766, Date: November 9, 2011\u003c/p\u003e","manuscriptTitle":"Variability in resistance training trajectories of breast cancer patients undergoing therapy","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-19 15:50:39","doi":"10.21203/rs.3.rs-4089501/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-08-18T07:43:03+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-16T22:28:04+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-08T20:50:42+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-06T16:02:12+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"163944755999419981594287405649253243529","date":"2024-07-31T04:11:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"41715429724998962527425075376804965655","date":"2024-07-26T19:59:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"292171467793793352852373809018199924748","date":"2024-07-26T19:09:49+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"270973026028042464289525774862005493598","date":"2024-07-26T06:11:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"182031471716368589592228803892842282481","date":"2024-05-13T13:14:09+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-04-27T04:04:37+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-04-21T13:16:28+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-03-15T04:48:57+00:00","index":"","fulltext":""},{"type":"submitted","content":"Supportive Care in Cancer","date":"2024-03-13T05:44:15+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"supportive-care-in-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jscc","sideBox":"Learn more about [Supportive Care in Cancer](https://www.springer.com/journal/520)","snPcode":"520","submissionUrl":"https://submission.nature.com/new-submission/520/3","title":"Supportive Care in Cancer","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"a7884e05-c881-482c-84a2-20ac9fa23bc1","owner":[],"postedDate":"March 19th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-12-16T16:04:01+00:00","versionOfRecord":{"articleIdentity":"rs-4089501","link":"https://doi.org/10.1007/s00520-024-09001-4","journal":{"identity":"supportive-care-in-cancer","isVorOnly":false,"title":"Supportive Care in Cancer"},"publishedOn":"2024-12-10 15:57:14","publishedOnDateReadable":"December 10th, 2024"},"versionCreatedAt":"2024-03-19 15:50:39","video":"","vorDoi":"10.1007/s00520-024-09001-4","vorDoiUrl":"https://doi.org/10.1007/s00520-024-09001-4","workflowStages":[]},"version":"v1","identity":"rs-4089501","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4089501","identity":"rs-4089501","version":["v1"]},"buildId":"pf3fE39SIOqb-0xH_OWvX","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}