The PRISMA flowchart shows the search results (Fig. 1). A total of 2959 publications were obtained from the bibliographic database searches. There were 1914 records identified from the Scopus database, 901 records from the PubMed database, 30 records from the international HTA database and 114 records from the National Health Service Economic Evaluation Database. A total of 561 duplicate records were removed before screening. After de-duplication, 2398 publications were included in the title and abstract screening. A total of 2341 publications were excluded during the title and abstract screening and 57 publications were included in the full-text review. The three main reasons for exclusion were the absence of a full health economic evaluation, the absence of simulation modelling and the absence of therapy response monitoring. Based on a full-text review, a total of 47 further publications were excluded. Reasons for exclusion are reported in Fig. 1. The number of publications included for data extraction and synthesis was ten [31,32,33,34,35,36,37,38,39,40]. Two studies were published in a scientific journal and as a full report within the UK National Institute for Health and Care Research HTA programme [31, 34, 39, 40]. For these studies, data were extracted based on the information available in all relevant reports, resulting in a final sample of eight unique studies [32,33,34,35,36,37,38,39]. For two studies, data were extracted only for the subgroup of patients with advanced disease who received systemic therapy [34, 37].

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) 2020 statement flowchart visualises the search results [23]. A total of eight unique studies, for which ten reports were consulted, were included for review. Note that reasons for exclusion during eligibility assessment are also included. iHTA international health technology assessment, NHS-EED National Health Service Economic Evaluation Database

Most studies were conducted in Europe [33,34,35, 37,38,39], one study in the USA [32], and one study in Canada [36]. All studies were published in a clinical journal. Most studies adopted a healthcare perspective [32, 34, 36,37,38,39], while only two Dutch studies considered a societal perspective covering potential productivity losses [33, 35]. One study studied patients with malignant lymphoma [33], four studies studied patients with advanced head and neck cancers [32, 35, 36, 39], one study studied patients with brain tumours [38], one study studied patients with non-small cell lung cancer [37], and one study studied patients with late-stage and persistent cervical cancer [34]. All studies studied responses to chemotherapy. Prescribed medication differed between the studies and was not always reported. Most studies considered first-line systemic therapy [32,33,34,35,36,37, 39], but also second-line therapy [37] and adjuvant therapy [38] were considered. The simulation time horizons adopted ranged from 1 year to lifetime. Three studies considered a lifetime horizon [34, 36, 39], two studies considered a fixed horizon of 5 years [33, 37], and one study considered a fixed horizon of 1 year [32]. The remaining two studies considered a horizon shorter than 1 year and adaptive to the therapy duration [35, 38]. A study-level overview is provided in Table 1.

All studies compared PET/CT using [1⁸F]Fluorodeoxyglucose (FDG) as PET-radiopharmaceutical with morphological imaging [33,34,35,36,37, 39], except for one study that considered stand-alone PET using [1⁸F]Fluoroethyl-L-tyrosine as a PET radiopharmaceutical [38]. One study did not report the PET radiopharmaceutical used [32]. In six studies, therapy response monitoring was scheduled within 3 months after the end of treatment [32, 34,35,36,37, 39]. In two studies, early therapy response monitoring was scheduled after two initial cycles of therapy [33, 38]. The performance of PET/CT was compared with contrast-enhanced CT [32, 34, 36, 37] or MRI [34, 35, 38]. One study included X-ray imaging as a relevant comparator [37]. Imaging test results were supported with histological proof in four studies [34, 35, 38, 39]. Histological proof was not reported in the other studies [32, 33, 36, 37]. One study positioned PET/CT as a replacement scan for morphological imaging [37], while the other studies positioned PET/CT as add-on scans [32,33,34,35,36, 38, 39]. As PET/CT is an expensive imaging modality, hospitals want to restrict its use as an add-on scan to patients who likely benefit from the imaging result. In general, add-on scans increase the sensitivity of an existing care pathway but potentially at the cost of specificity [11]. Depending on the positioning of PET/CT within the care pathway, diagnostic performance measures of sensitivity, specificity, positive predictive value or negative predictive value were alternately used in the simulation models.

Imaging test results impact therapy selection and further patient management. Four studies reported imaging-based stratification of patients to subsequent therapies [34, 36, 37, 39]. These studies distinguished curatively and palliatively intended therapy options and best supportive care. Moreover, four studies reported the avoidance of overtreatment [32, 33, 35, 38]. In these studies, patients were prevented from receiving futile or unnecessary therapy. Furthermore, five studies estimated patients’ survival after sequential imaging and therapy [33, 34, 36,