Twelve-Weeks of Virtual and In-Person Exercise Training Improve Static and Dynamic Balance in Female Cancer Patients

Background: Balance is important for maintaining activities of daily living and functional independence. Whether balance is improved in cancer patients after tness training is unclear. The purpose of this study was to determine if 12-weeks of exercise improves balance and posture in female cancer patients. Methods: Thirty-eight female cancer patients were enrolled. Of the 16 patients who completed the program (mean age±SD: 65±10 years), 12 participants were novice and 4 were experienced exercisers. Six-weeks of exercise sessions were provided in-person and the remaining sessions were virtually delivered. The American College of Sports Medicine’s exercise recommendations were followed. Novice exercises received 36, 90-minute exercise sessions (3x/week) and experienced exercisers had 24, 90-minute sessions (2x/week). Posture was measured using the plumb line method and overhead squat test; balance was measured using the unipedal single leg stance and limits of stability (LOS). Body composition, cardiorespiratory tness, muscular strength, and exibility were also measured. Two-way repeated measures ANOVAs and Bonferroni’s multiple comparison tests were used to determine signicant differences. Results: Balance on the left leg (eyes closed) and LOS with leftward excursion were signicantly improved in experienced exercisers (P=0.0029), but not in the novice group (P=0.0013). Qualitative data showed that experienced exercisers had improved static and dynamic postural alignment of the lower body. Conclusions: While cardiorespiratory and muscular tness signicantly improved in the novice group, these patients did not show balance and postural improvements. Only the experienced exercisers had signicant improvements in static and dynamic balance and lower body posture.


Introduction
Balance, posture and muscular performance are important for maintaining functional independence and activities of daily living (ADL) [1][2][3]. Posture is the orientation of a body segment relative to a gravitational vector, and it is affected by the downward pull of gravity [4]. Balance is related to the inertial forces acting on the body, and inertial characteristics of individual body segments [4], and more simply, the ability to prevent falls [5]. Physical tness not only affects the ability to accomplish ADL, but it relates to psychosocial functioning as impaired balance could lead to social isolation, activity avoidance and fear of falls [6]. As such, the ability to accomplish ADL is positively associated with QOL [7].
Cancer treatments can affect patients' quality of life (QOL) and hinder their ability to engage in physical activity [8]. Balance may be aggravated by a lack of muscle conditioning and exibility, poor posture maintenance and neurotoxic chemotherapeutic agents of therapy [9,10]. In the case of breast cancer, treatments such as surgery can result in dysfunction and pain in the upper limbs, ultimately limiting range of motion (ROM) in exion, abduction, and external rotation shoulder movements [11]. Surgery or radiation can cause brosis or densi cation of fascial structures that decrease limb mobility [11]. A decrease in ROM and pain can result in failure to be active, further provoking poor muscular function, which can negatively impact posture [11]. Proper sagittal spinal alignment and muscle strength improves balance, and reduces the risk of falls [2]. Since the body can be interpreted as a kinetic chain, changes in the upper body and spine alignment contribute to deviations in the lower body, which could impair balance [12].
A balance-based exercise intervention has shown to partially improve postural instability in cancer patients with severe chemotherapy-induced peripheral neuropathy (CIPN) [13]. However, the impact of general tness programs on balance is unclear. Exercise programs which follow the American College of Sports Medicine's (ACSM) recommendations for cancer patients, which target cardiovascular and muscular tness, and exibility improve aerobic and muscular tness [14] but its impact on balance and posture are not clear. The impact of exercise on posture and balance in cancer patients has not been comprehensively studied and knowledge is often derived from diabetic neuropathy studies [13]. Thus, the purpose of this study was to evaluate the effect of 12-weeks of exercise training on posture and balance in novice and experienced female exercisers who were diagnosed with cancer.

Design and Procedures
Thirty-eight female patients who completed cancer treatment were enrolled. Patients were recruited through a state-wide referral network comprised of oncologists, radiologists, and surgeons. All participants received exercise clearance from their oncology-related medical provider and medical histories were provided. Inclusion criteria included being female, having been diagnosed with cancer, having completed clinical cancer treatments at least 3 months previously, being ambulatory, over the age of 18 years, received exercise clearance from their provider, literacy in English, and ability to attend exercise sessions for 12 weeks. Prior to participation, patients provided their verbal and written consent.
Research activities were approved by the Institutional Review Board (IRB, #2018 − 00167).
Initial and nal tness testing took place in-person; nal assessments were conducted during the pandemic using additional safety precautions as approved by the IRB. At the mid-point of the exercise intervention, the COVID-19 pandemic emerged, precluding in-person, individualized exercise training. All participants were given the option to continue their program virtually during the quarantine. Of the 38 patients, 16 completed the intervention (mean age±SD: 65±10 years). Twelve of these patients were new to the program and these novice exercisers received 3 exercise sessions per week (36 sessions total). Four patients were experienced exercisers, having completed a 12-week face-to-face exercise program prior to their current enrollment. In the current study, they received 2 exercise sessions per week (24 sessions total). Participants included those diagnosed with breast cancer (invasive ductal carcinoma and invasive lobular carcinoma in either the right or left breast, n = 14), endometrial carcinoma (n = 1) and Non-Hodgkin's lymphoma (n = 1). Cancer treatments completed by breast cancer patients included breast lumpectomy or mastectomy (right, left or both sides, n = 13), conserving therapy (n = 1), reconstruction (n = 3), radiation (n = 7), chemotherapy (n = 6), and hormonal therapy (utilizing Herceptin, Anastrozole and/or Tamoxifen, n = 11).

Balance measures
Unipedal single leg stance (SLS) was used to measure static balance. Patients were instructed to balance on each bare foot, beginning with their dominant foot for as long as possible up to a maximum time of 45 seconds. This was performed with and without visual feedback, eyes open (EO) and eyes closed (EC), on both right (R) and left (L) feet. The best score out of 3 trials was recorded for EOR, EOL, ECR and ECL.

Postural measures
All postural measurements were qualitatively assessed. Subjective descriptions such as "mild" or "signi cant" variations in posture have used to describe the severity of an abnormality [15]. However, the use of this method is limited by subjective observations and therefore, the presence or absence of a deviation was recorded and frequencies of postural deviations were analyzed in each group.
Overhead Squat. The National Academy of Sports Medicine's (NASM) overhead squat assessment was used to assess muscle imbalance and postural deviations during a dynamic squat [16]. Patients were asked to stand with their feet-distance apart, toes pointing forward, heels on the oor and arms extended overhead. Then, they were asked to slowly squat down to sit in a chair while evaluators noted upper and lower body anatomical deviations (i.e., forward head/arms/torso, knees and toes pointing inward/outward).
Plumb line Method. This method was used to assess posture deviations in a static position in anterior, sagittal and posterior views. Patients were asked to stand comfortably in the anatomical position with their feet shoulder width apart and arms hanging to the side while evaluators noted prominent deviations. In the sagittal view, the presence or absence of forward head, forward lean, scapular protraction/retraction, kyphosis, lordosis and posterior or anterior pelvic tilt was recorded. In the anterior view, the following deviations were notated: head tilt or rotation, shoulder elevation, high hip, valgus or varus knee, pronation or supination of feet and inward or outward pointing toes. The posterior view was used to detect weight shifts, hyperextension of the knees or elevation of hips.
Cardiorespiratory, body composition, exibility, muscular tness measures Prior to physical tness testing, resting vital measurements were used to verify that all subjects had normal blood pressure, oxygen saturation and sinus rhythm. Body composition, muscular strength, exibility and cardiovascular endurance were measured pre-and post-intervention. These measures helped to identify baseline tness characteristics as well as show changes in tness. Body composition was measured with bioimpedance analysis (Omron, HMF306), waist circumference measurements, waistto-hip ratios and body mass index. Lower body muscular strength was assessed using 1-repetition maximum (1-RM) tests using the horizontal leg press, seated leg extension and seated leg curl exercises. In cases where the 1-RM could not be achieved, a prediction equation was used to estimate 1-RM using the maximum weight lifted for 10 or less repetitions [17]. where nal speed and grade were used to estimate and maximal oxygen uptake (VO 2 peak) [19]. The test was terminated upon volitional fatigue, the patient's request to stop or if contraindications to exercise were observed.

Exercise Training
Fifty percent of the exercise intervention was provided in person with the remaining 50% being provided virtually due to COVID-19 restrictions. The in-person training was led by an exercise specialist knowledgeable in cancer exercise rehabilitation principles. Exercise specialists were trained using a specialized curriculum [20]. Patients were led through 90-minute exercise sessions and at least 1 rest day was placed between training sessions. Speci cally, patients completed 30 minutes of cardiovascular training and 5-7 resistance training exercises at a workload of 40-85% of the subject's 1-repetition maximum (1-RM). Heart rate reserve (40-85% of HRR) [21] and Borg's Rate of Perceived Exertion (RPE) were used to determine indices of physiological strains during cardiovascular training [22]. Heart rate (HR) was monitored throughout the session using a chest-based device that wirelessly connected to an external tablet (Polar H10, USA). Target RPE was 3-6 (on a scale of 0.5 to 10, where 0.5 corresponds to resting and 10 corresponds to maximal exertion). RPE was a necessary tool for patients who use pharmacological drugs that lower HR such as β-blockers, non-dihydropyridine calcium channel blockers and ivabradine [23]. RPE was reported after each exercise bout and helped to identify true physiological strain independent of HR response to exercise.
Balance was trained using foam pads, BOSU balls, stairs or other techniques to work on strengthening leg, foot and back muscles. Exercises that used tandem or split stances in resistance training exercises were used to train balance. With regard to postural training, exercise specialists used results of the NASM overhead squat assessment to identify speci c muscles to stretch or strengthen with the goal of improving posture [24]. All exercise programs incorporated exercise training principles and followed recommendations of the ACSM for patients diagnosed with cancer, i.e., aerobic, resistance training, balance and exibility training [19,25]. Logbooks were used to document exercises and assure delity to the protocol.
Virtual programs followed the same guidelines listed above. Trainers provided the exercise program through email or Google Docs and patients reported their progress in the same manner or through phone or text messaging. Trainers utilized exercise equipment the patient had at home or outdoor activities (i.e., open water swimming, walking in the neighborhood) to create the exercise program. Trainers utilized online resources (i.e., videos which demonstrated the exercise) or shared videos of themselves performing the exercise. Trainers then relayed patient progress to administrators, and this correspondence was used to assure delity to the protocol.

Statistical analyses
Quantitative tness data were analyzed with 2-way repeated measures ANOVAs with factors Time (pre, post) x Group (novice, experienced). Bonferroni's posthoc multiple comparisons tests were used to detect differences between groups pre-to post-intervention; signi cance was set at P < .05 (GraphPad Prism, version 8.0.0 for Windows, San Diego, CA). Pilot data were used in power analyses to show that with 9 subjects, there would be a 95% chance of detecting improved static balance with a mean improvement of 5 seconds and incorporating 10 subjects would result in a 95% chance of detecting improved dynamic balance (power, 1-β = 0.8, one-tailed, SPSS ver. 27). Qualitative data for dynamic and static posture were organized into tables to show the frequencies of which a postural deviation was observed.

Results
The exercise intervention was effective in improving overall tness. Pre-and post-tness data are provided in Table 1   Results of the static plumb line and dynamic NASM overhead squat assessments are shown in Tables 2  and 3, respectively. The frequency of each deviation was summed. In the novice group, knee and foot positioning (outward motion) and weight shifting worsened as measured with the dynamic postural assessment. A decline in foot positioning was also detected with the static postural assessment. With static posture, lordosis, elevation of the right shoulder, hyperextension of the knee was exaggerated after the intervention. Notably, there was improved knee and head positioning, reduced weight shifting and improved elevation of the left shoulder with static posture. Additionally, the novice group had a reduced incidence of foot attening during the NASM squat test.

Discussion
Altogether, the results show that female cancer patients were able to improve their cardiorespiratory tness, waist circumference, and exibility as detected with a signi cant main effect of time. It is surprising that static balance was not signi cantly improved in novice exercisers, but was improved in the experienced group on their non-dominant foot. This improvement in balance occurred without concomitant gains in hamstring exibility or lower body muscular strength. This is unexpected as hipmobility is positively associated with balance in older adults [26] and previous studies showed that strength is associated with balance [1,27]. Differences in the study methodologies could account for these differences (i.e., different markers of exibility or strength). Similarly, experienced exercisers showed a signi cant improvement in leftward excursion of their center of gravity during dynamic balance testing, representing enhanced dynamic balance in the non-dominant direction. The experienced exercise group fell in the 88 th percentile at baseline, which allowed for greater gains in dynamic balance to be achieved. Whether improvements in balance would have been detected in the novice group if all exercise sessions had face-to-face exercise leadership is unknown.
Since signi cant improvements in balance were observed in the returning exercisers, this suggests that there may be a timeline of improvement associated with enhanced balance in this population. New exercisers had signi cant improvements in VO 2 peak, muscular strength and hamstring exibility, while these improvements were not signi cantly increased in returning exercises. This may indicate that balance development occurs after cardiorespiratory and muscular tness gains. Also, the experienced exercisers accomplished a 45-second hold (which was the maximum time) during the baseline SLS test with visual feedback on both right and left feet, but this was not shown in the novice group. This suggests that patients may have required a baseline tness to have signi cant gains in balance. Another possibility is that patients with ample exercise experience were able to achieve signi cant improvements in balance because they were knowledgeable about using proper training techniques. This may be especially important in the current study because one-half of the intervention was delivered virtually.
It is striking that signi cant improvements in balance and tness were achieved with the virtual exercise intervention because a systemic review and meta-analysis showed that supervised training programs have superior effects on tness and QOL compared to home-based regimens [28]. This approach is bene cial because it reduces sta ng resources associated with the exercise intervention. In fact, providing patients with in-person training followed by independent exercise training may be an effective approach for enhancing balance in the cancer patient population, especially in individuals with exercise experience. Whether larger gains in balance could have been achieved with face-to-face exercise leadership for 12 weeks in experienced exercisers remains to be determined.
The observation that knee alignment during the NASM squat test worsened in the novice group suggests that this group may have bene tted from in-person training to ensure correct posture and alignment were used during exercise training. It is notable that in 2 of 3 novice exercises, there was a reduction of foot attening during dynamic movement. Exercises targeting lower leg and feet musculature have been shown to improve medial longitudinal arch and balance [29], but in the current study this did not signi cantly impact balance. The experienced group, although comprised of a small sample, had consistent, positive lower body postural changes as a result of the intervention. However, upper body static posture did not have the same result. This suggests that the experienced group was capable of training with proper lower body posture, while additional oversight may have been bene cial when targeting upper body musculature. Still, results suggest that using a combination of in-person and virtual exercise programming may provide more postural bene t for the experienced exerciser when compared to an individual with little exercise experience. Overall, results of the current study correspond to ndings that computer-feedback balance training improves balance outcomes in patients with CIPN [30], and aerobic, strength and sensorimotor training improves physical functioning and independence in patients with CIPN [31].
ADL require the use of multiple sensory inputs including visual, vestibular, and somatosensory systems; and when these systems are interrupted, they impair dynamic balance [6]. These systems are utilized more when performing dynamic balance assessments, whereas in static balance they are utilized to a lesser degree. Dynamic posture may be better suited to show direct implications on ADL and its related movements, such as rising from a chair and walking or performing manual tasks while standing [32]. Since static and dynamic balance are achieved through different sensory processes, both parameters should be taken into account when assessing balance.

Limitations
The main limitation in this study is due to pandemic-related events. Approximately 58% of the patients discontinued the program due to the mandatory quarantine and transition to a virtual exercise format.
Patients withdrew because they felt they would bene t from an in-person program rather than a virtual intervention. Patients with reduced mobility sought in-person exercise training with specialized exercise equipment, and/or others felt that they could not be accountable for their exercise program. Thus, the patient population in the current study represents individuals who felt con dent that they could independently complete the exercise intervention. The current study did not measure symptoms of CIPN; future research should quantify CIPN severity as it may be related to improvements in balance and posture. Even when faced with these limitations, signi cant improvements in balance were detected, indicating that the approach was effective in improving balance.
Analyzing postural data is di cult due to the qualitative and subject nature of the static and dynamic postural assessments. A lack of inter-rater reliability may have skewed the postural data. However, the identi cation of the presence or absence of a postural deviation minimized the subjective nature of the rating. A standardized and objective measure would improve future studies evaluating changes in posture. Additional work is necessary to validate the ndings of the current study.

Conclusions
The intervention was effective in improving whole-body tness in both novice and experienced groups.
The experienced group had signi cant gains in static and dynamic balance and improved lower body posture. While the novice group had signi cant improvement in cardiorespiratory tness and exibility, they did not show signi cant improvements in balance.

Declarations
Funding: None to report. The authors report no involvement in the research by the sponsor that could have in uenced the outcome of the work.