Log in to the Web of Science database platform (http://apps.webofKnowledge.com/), select the core collection database, and refer to the existing literature and the procedures and standards of scientometric analysis to formulate a literature retrieval strategy (31–34), as shown in Figure 1 .

Entered the count items, which meet the inclusion criteria, into Excel2021, such as publication time, citation frequency, and number of articles published by different subjects (e.g., authors, institutions, countries/region, et al.). Used Oringin 2021 and PyCharm Community Edition 2022.1.3 to map the publication time and citation frequency characteristics of publications, along with the development trend map in the next 5 years. The H-index obtained from the WoSCC webpage is used to assess the scientific impact of academic outputs of different subjects (35).Then import the publication information downloaded from the website into CiteSpace6.3R1, VOSviwer1.6.20 or the Bibliometric Analysis Platform (http://biblimetric.com). To understand the development status of the discipline by drawing the scientific research co-authorship map of countries/regions, institutions and authors, to explore the research hotspots and development trends of the subject by drawing the co-occurrence keyword map and the timeline diagram of co-occurrence keyword clustering, to analyze the high-impact journals in the subject area by drawing the co-cited journal map, and to understand the research frontier and its evolution path of the subject area by drawing the co-cited reference cluster map. We draw the scientific map, and the specific parameters are shown in the parameter information in the upper left corner of the picture.

In the scientific knowledge graph, each node represents an element, representing the country/region, institution, author, co-cited author, co-cited journal, co-cited reference, and co-occurrence keyword. There is a positive correlation between node size and the number of published papers, co-occurrence frequency or co-citation frequency. The node color represents the time when it first appears, and the evolution process from cool color to warm color represents the development path of its appearance time from far to near. The outermost purple circle of a node represents the betweeness centrality≥0.1, which is generally considered to be a key node in its domain, that is, it has an important “bridge” role. The annual rings of a node represent its citation history, the size of the annual rings reflects the number of times it has been cited, the color of the annual rings represents the citation time, and the thickness of the annual rings is proportional to the number of citations in the corresponding time partition. The diameter of the connection between nodes is positively correlated with the number of co-authored papers, co-occurrence frequency or co-citation frequency. The color of the connection represents the time when it was first co-authored, co-appeared, or co-cited, and the evolution process of the network connection from cool to warm represents the evolution of the research field from far to near in time (31, 32).

In a clustered view, the larger the number of members contained in a clustered cluster, the smaller its number. (1) The larger the Modularity value, the better the clustering effect, and when Q>0.3, the structure of each clustered cluster is significant. (2) The larger the Silhouette value, the higher the similarity of the members within the clustering cluster (when there are few members within the cluster, its reliability will decrease); when the Silhouette > 0.7, it indicates that the reliability of the clustering results is higher, when it is 0.5~0.7, the clustering results are considered reasonable, and the closer the Silhouette value is to 1, the higher the homogeneity within the cluster. (3) The more nodes contained in the clustered cluster, the more important the research field involved in the clustered cluster, and its time span is positively correlated with the duration of its research popularity. (4) The more burst nodes contained in the clustered cluster, the more active the research content involved in this field is, and it also suggests that it may be an emerging trend in the field of research (33, 34).

In the emergent map, the burst node is filled with red in the corresponding burst year, and its time span is positively correlated with the duration of the research enthusiasm of the node. In the timeline view, nodes in the same clustered cluster are placed on the same horizontal line, and the node occurrence time is placed at the top of the graph, and the nodes on the right are more recent. It is worth noting that burst detection is often used to find nodes with “bridge” role in different clustered clusters in cluster maps and time views (36).

We found a total of 512 relevant publications, which were cited a total of 12013 times, with an h-index of 57. The distribution characteristics of their publication time and citation times are shown in Figure 2A , and predict their development trends over the next 5 years as shown in Figures 2B, C .

There are a total of 17091 co-cited references, of which only 27 are cited more than or equal to 28 times. The clustering map of the co-cited references was well structured, and the internal consistency of each cluster was high (Q=0.8634, S=0.9471, Figure 3A ). The 14 clusters the evolution process of cluster names from cool to warm colors represents the evolution path of the research frontier in this discipline from far to near in time) were constructed into six main research trends( Figure 3B ). The 1^st item: identified in 1997 and mainly involved “following stem cell transplantation” (#12). The 2^nd item: identified in 2015 as a reference to “life” (#26). It continues to be the most active research area throughout the development of the subject area. The 3^rd item: identified separately in 2012- 2019. It mainly involves the intervention research on cancer fatigue and sleep quality in cancer patients, including exercise level, venue, program, et al. (#0, #1, #3, #6, #8, #9), and is also the most popular and cutting-edge research content in this field. The 4^th item: identified in 2008-2010. It primarily deals with the clinical practice of exercise therapy (#4, #5). The 5^th item: identified separately in 2005-2007. It primarily involves studies of nursing intervention programs in exercise therapy (#7, #11). The 6^th item: Identified in 1999, mainly related to “multiple myeloma”. Moreover, the emergent map of co-cited reference clustering suggests ( Supplementary Figure 1 ): Campbell et al. (37) published in Med Sci Sports Exerc in 2019 had the highest number of co-citations (research popularity), and Mustian et al. (38) published in JAMA Oncol in 2017 was the most cited and pointed out that the combination of exercise intervention and psychotherapy can significantly reduce chronic renal failure during and after cancer treatment, and recommended that clinicians should use it as the first choice for first-line treatment of cancer. In addition, the current research frontiers in this field also focus on the relationship between physical activity, sedentary behavior, and cancer prevention and control (39), as well as the active assessment and management of the medical and psychosocial needs of cancer patients at all stages (40).

The high-frequency co-occurrence keywords of the subject area reflect its main research hotspots. A total of 508 keywords were extracted by CiteSpace software ( Figure 4A ), of which 208 keywords appeared at least 5 times, and the top 10 keywords in terms of frequency were: quality of life (253 times), physical activity (158 times), and breast cancer (120 times), exercise (112 times), fatigue (92 times), survivors (75 times), women (62 times), depression (58 times), Prevalence (57 times), health (56 times), sleep (56times). Therefore, we believe that these co-occurrence keywords may be the research hotspots in the field of tumor-related sleep-wake disorders. In addition, Figure 4A also suggests that the top 10 co-occurrence keywords of subject influence are: breast cancer (betweeness centrality: 0.21), health(0.21), aerobic exercise (0.2), adults (0.17), chemotherapy (0.16), cancer(0.15), quality of life (0.13), exercise (0.12), fatigue (0.11), cancer-related fatigue (0.11), children (0.11). In the timeline diagram of co-occurrence keyword clustering ( Figure 4B ), the transformation process of cluster names from cool to warm colors reflects the development trend of research hotspots in this discipline. The co-occurrence keyword emergence map suggests that “adjuvant chemotherapy” has been the long-term concern of scholars (1995-2012), and “fatigue” has been the most popular (2012-2015). However, at present, there is no co-occurrence keyword with the highest research enthusiasm and active state.

Among the 97 journals that published papers, only 24 journals published 5 papers or more, with Support Care Cance (71 papers), Cancer Nurs (42 papers), Psycho-Oncol (30 papers) publishing the largest number of papers, J Clin Oncol (366 papers), Support Care Cancer (320papers), and Cancer (301 papers) was the most frequently cited, with Cancer (betweeness centrality: 0.1), Brit J Cancer (0.09) and Cancer Nurs (0.08) having the highest impact. The emergent map of the co-cited journals ( Supplementary Figure 3 ) showed that Cancer had the highest emergence intensity (10.47) and the longest duration (1992-2013). Cancers also had the highest emergence-intensity (9.15), with Medicine, JAMA Netw Open, Nutrients, Int J Env Res Pub He, SCI Rep-UK, and Jnci Cancer Spect remain the most active journals currently (2021-2024).

Through the analysis of the co-cited reference network, our study identified the following research frontiers in this discipline: the efficacy goals of exercise therapy in the adjuvant treatment of patients after tumor resection (43), physical behavior and activity levels (44, 45), exercise levels (45, 46), exercise patterns (46), exercise sites (47), treatment regimens (48–50), and the clinical efficacy and mechanism of exercise therapy in improving sleep therapy and living standards in chemotherapy patients (51). Typical representative studies suggest: cancer patients should be helped to establish healthy eating behaviors and develop individualized exercise behavior and exercise level intervention strategies (44, 45). Although moderate-intensity walking is effective in improving sleep in cancer patients (45), aerobic exercise combined with resistance exercise may be more clinically effective (46), and exercise therapy interventions are best tailored to the patient’s physical impairments at baseline, such as: loss of muscle strength, cardiopulmonary decline, or sleep disturbances, et al (48). In addition, insomnia is a predictor of cancer progression and quality of life, and vestibular stimulation is particularly appropriate for the treatment of its accompanying symptoms (49). Menopausal symptoms are the strongest predictors of menopausal symptoms in patients with gynecologic cancers, and assessment of menopausal symptoms and sleep quality during and after cancer treatment, timely intervention (51), reduction of exercise-induced discomfort, and control of insomnia can effectively improve their health-related quality of life (43). We also found that “life” is the research frontier which runs through the subject area from beginning to end, and it is still active. Recent studies have concluded that the quality of life of cancer patients is positively correlated with their income and physical activity level (52), and the symptom management system (SMILE) can improve their life and have a positive clinical effect on fatigue and sleep disturbance after adjuvant or palliative chemotherapy (53). In short, mastering the above research frontiers in this discipline will help clinicians and relevant health care personnel to formulate targeted exercise therapy intervention programs and measures according to patients’ personality characteristics, so as to scientifically guide and supervise patients to implement exercise therapy scientifically and efficiently, which will be of great significance for managing patient compliance and promoting the healthy development of the discipline.

In this study, we identified a total of 111 keywords that appeared 5 times or more. The results showed that the co-occurrence keywords, such as: quality of life, physical activity, breast cancer, exercise, fatigue, and survivors, appeared most frequently. So we thought that thy may have been the research hotspots in this subject area. From the perspective of the connection strength and node size of co-occurrence keywords in the scientific knowledge graph, we also found that breast cancer, quality of life, exercise, physical activity, fatigue, chemotherapy, health, adjuvant chemotherapy, etc. have had significant academic influence in this discipline. Therefore, we believe that they deserve special attention and in-depth study. A representative study suggests that symptom clusters in Asian-American breast cancer survivors may improve with the duration of individual/group/supportive technology interventions (54). High-intensity aerobic interval training (HIIT) combined with resistance training (RES) has been shown to improve adverse effects such as depression, daytime sleepiness, and insomnia in patients undergoing radiotherapy and chemotherapy for rectal cancer (55). Regular endurance training combined with resistance training can improve mental health, sleep, quality of life, and physical health in cancer survivors (56). Providing telehealth cognitive and behavioral training for insomnia may also help improve sleep quality in cancer survivors in rural or economically disadvantaged areas (57). Moderate- to high-intensity physical activity may be of greater benefit to patients with bone metastases. Reducing sedentary behavior may be a key goal for patients with a history of fractures (58). Oncology nurses play a key role in providing education about the benefits of exercise, overcoming barriers to physical activity, and timely referral (59). In addition, it can be seen from the timeline diagram of co-occurrence keyword clustering that the transformation process of cluster names from cold to warm reflects the development trend of research hotspots in this discipline from far to near in time, and also suggests that the research topics related to “radiotherapy-related fatigue”, “controlled trial” and “nutritional state metabolism” are still the current research hotspots. They may be involved in survivors, prevalence, breast cancer, physical activity, sleep, fatigue, health, depression, et al. This is also consistent with the visualization results of the co-occurrence keyword scientific knowledge graph.

Although our study elaborates the research status of exercise intervention in tumor-related sleep-wake disorder through bibliometric and visual analysis, and provides a relatively objective and detailed basis, there are still some unavoidable limitations. First of all, the literature data of this study came from the WoSCC database, and there are still some excellent publications that have not been included. Second, our study excluded non-English publications and papers other than original research.