Current evidence and future perspectives in the treatment of head and neck vagal paragangliomas: a systematic review Authors: Lara Valentina Comini, Andrea Rampi, Emilio Salerno, Paola Lovino Camerino, Alex Finco, Gabriele Redaelli, Davide Di Santo, Stefano Bondi Published: 2026 ABSTRACT AbstractBackgroundVagal paragangliomas (VPGs) are rare neuroendocrine tumors, typically benign but with unpredictable growth and local invasiveness. Their management is debated, particularly in asymptomatic or elderly patients, due to the lack of clear predictive indicators for progression and treatment response. This systematic review aimed to assess the clinical features and treatment outcomes of VPGs.MethodsWe conducted a systematic review of the literature, screening studies at multiple levels. A total of 41 articles were included, comprising data from 501 patients. Treatments were categorized as surgery, radiotherapy (RT), both, or observation (wait-and-scan). Due to study heterogeneity and generally low evidence levels, a statistical comparison between treatment modalities was not feasible.ResultsAmong 501 patients, 325 underwent surgery, 101 received RT, 9 had combined treatment, and 66 were managed conservatively. Surgical resection achieved effective tumor removal and long-term control but was almost always associated with vagal nerve injury, leading to dysphonia and dysphagia. There was a 2.6-fold increase in postoperative impairment of at least one additional cranial nerve (IX, XII) or the sympathetic chain compared to preoperative status. RT and stereotactic radiosurgery (SRS) offered high tumor control rates with generally mild side effects, though long-term radiation risks—especially in younger patients—remain a concern. These modalities were often preferred for small, non-secreting tumors with preserved nerve function. The wait-and-scan approach was mainly used in elderly patients with asymptomatic or slow-growing tumors. No definitive predictors of tumor growth or treatment complications emerged from the literature.ConclusionsSurgery offers curative resection but carries a high risk of cranial nerve morbidity. RT and SRS are effective, less invasive alternatives with favorable safety profiles, particularly suitable for small, stable tumors. While observation remains appropriate in select cases, emerging options such as proton therapy, partial resection, and theranostics require further research to clarify their roles. The overall quality of evidence remains low, highlighting the need for prospective, comparative studies.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13643-026-03212-y.Keywords: Head and neck surgery, Vagal paragangliomas, Radiotherapy, Systematic review Background Paragangliomas (PGs) are rare neuroendocrine tumors that develop in the paraganglia, small organs mainly consisting of specialized neural crest-derived cells organized in a diffuse network embedded in the autonomic nervous system [1]. At the cellular level, PGs are indistinguishable from pheochromocytomas and may arise anywhere along the parasympathetic and sympathetic chains [1]. Due to the significant presence of autonomic nervous structures in the head and neck, a significant percentage of PGs, especially those of parasympathetic origin, arise at this site [2, 3]. They account for < 1% of all head and neck neoplasms, but are relevant due to their peculiar anatomical sites, the possibility of local infiltration, and the relationship with genetic disorders [2]. Approximately 60% of all head and neck PGs (HNPGs) are located in the carotid body (CBTs), followed by the jugular bulb (with possible tympanic extension) and along the vagus nerve (15–20%) [4]. Vagal paragangliomas (VPGs) often present as asymptomatic slow-growing lesions of the neck, while nerve impairment can manifest with symptoms such as dysphonia or dysphagia; injury to other cranial nerves may also occur, especially when the tumor compresses on or invades nearby structures [3, 5]. The increasing awareness of its possible genetic basis, such as mutations in succinate dehydrogenase subunit genes (SDHx genes), has revealed the hereditary origin of many VPGs, with improvements in patient screening [2, 4]. Nonetheless, the choice of the most appropriate treatment remains challenging due to the need to counterbalance the benefits and hazards of surgery, radiotherapy (RT), and watchful waiting. The present paper offers a systematic review of VPGs, aiming to provide an up-to-date summary of possible management strategies. Methods Search strategy According to the PRISMA reporting requirements, we performed a comprehensive search on PubMed, Embase, and the Cochrane Library databases for studies including original data regarding HNPGs on March 5, 2024, and it was set to automatically update until December 2024. Supplementary Table 1 shows the datasets, the string used for research, and the number of items retrieved from each database; Supplementary Table 2 provides the PRISMA checklist [6]. Following exclusion of duplicates, four investigators (L.V.C., P.L.C., A.F., and E.S.) assessed the titles and abstracts to eliminate publications that were clearly not relevant. The full texts of the remaining articles were evaluated in duplicate by two authors (L.V.C. and E.S.) for additional consideration. During the abstract evaluation phase, we included all research considered eligible by at least one rater. Disagreements were settled during the full-text review stage through consensus. In the case of multiple eligible articles from the same institute, only the most recent one was included to avoid duplicates or overlapping cases. At first, papers reporting all types of HNPGs were included (carotid, jugular, jugulotympanic, and vagal) to avoid the exclusion of significant cases, while a successive full-text reading led to the exclusion of non-vagal PGs; both benign and malignant forms were considered. A publication was included when it fulfilled the following criteria: (1) written in English; (2) patients were > 18 years; (3) had a sample size of at least 5. The latter requirement was set to include centers with higher experience, even in other subtypes of HNPGs, and avoid reports from inexperienced surgeons that could alter the results (such as overestimating treatment complications). Papers were excluded when (1) reported non-head and neck VPGs; (2) were meta-analyses, systematic and narrative reviews (but a hand-check for other potentially relevant studies was performed); (3) there was intracranial extension of the neoplasm (corresponding to stage III VPGs according to Browne et al. [7]). Data extraction The following information was recorded for each article: year of publication, time period studied, number of cases, nature of the study (prospective/retrospective, mono-/multicenter), age range, mean age, gender, tumor characteristics (histological type, single or multiple, primary or recurrent, side, size), clinical presentation (including nerves involved at presentation), treatment (wait-and-scan, surgery eventually followed by RT, RT alone, eventual preoperative embolization), treatment efficacy and complications (such as neural impairment of vagal, cranial VII, IX, XI, and XII nerves, and sympathetic chain and major surgical injuries), and follow-up (need for additional treatment, disease-free survival, and disease-specific survival). When a patient suffered from multiple PGs, most data were not split for vagal and non-vagal tumors, and therefore all were included. The National Heart, Lung, and Blood Institute Study Quality Assessment Tools (NHI–SQAT) were used to evaluate the studies for methodological bias and quality [8]. Using a methodology consolidated in previous mixed-level of evidence reviews, items were rated as good if they fulfilled at least 80% of the NHI-SQAT requirements, fair if they fulfilled between 50 and 80% of the items, and poor if they fulfilled less than 50% of the items. Results Literature research, risk of bias and description of the studies The PRISMA flow-chart of the study is summarized in Fig. 1. The database search on Embase, Cochrane, and PubMed retrieved a total of 1866 papers, which were reduced to 1597 after removal of duplicates (one paper was excluded due to publication withdrawal). The screening of the abstracts led to the removal of 1474 studies, mainly due to non-relevance, language, or different histology. As a result, a total of 150 articles underwent full-text examination; based on the abovementioned inclusion/exclusion criteria, 41 articles were included in the present review. PRISMA 2020 flow diagram for updated systematic reviews which included searches of databases and registers The design of the studies was mostly monocentric retrospective (82.9% of articles), apart from 4 multicentric [9–12] and 3 prospective designs [13–15]. The NHI-SQAT assessment classified 24% of studies as good, 20% as fair, and 56% as poor. We did not exclude any studies depending on the NHI-SQAT assessment, and their evidence was not systematically less considered. As a general trend, thus, higher-quality studies provided a more complete description and this may have led to a greater contribution to data for the review. Table 1 provides the details for each article. Table 1Description of the studies included. Quality assessment: + (good ≥ 80%), −/+ fair (80–50%), − poor (≤ 50%). *Multicentric, **prospectiveAuthorsYear of publicationStudy periodCountryQuality assessmentTotal n. of patientsn. of patients includedLi, L.et al. [16]20232010–2019Asia − 141Tamaki, A. et al.* [9]20222011–2020USA −/+ 12524Guha, A.; Chovanec, M.** [13]20222016–2020Europe + 309Contrera, K.J. et al. [17]20201990–2010USA −/+ 17320Singh, S. et al. [18]20192009–2013Asia − 546McCrary, H.C. et al. [19]20191963–2018USA − 447Papaspyrou, K. et al. [20]20091989–2007Europe −/+ 12012Mendenhall, W.M. et al. [21]20191968–2016USA − 14936Dupin, C. et al. [22]20141990–2009Europe − 6617Michelozzi, C. et al. [23]20162005–2012Europe −/+ 297Persky, M.S. et al.* [10]20021990–2000USA − 475Paris, J. et al. [24]20061990–2004Europe − 3211Konefal, J.B. et al. [25]19871961–1983USA − 332Hinerman, R.W. et al. [26]20011968–1998USA −/+ 7110Tasar, M.; Yetiser, S. [27]20041992–2002Europe − 172Siedek, V. et al.** [14]20092009Europe − 61Antonitsis, P. et al. [28]20061990–2004Europe − 131Martínez-Galdámez, M. et al. [29]20112008–2010Europe − 61Piazza, P. et al.* [11]20131988–2010Europe − 207Künzel, J. et al. [30]20142000–2012Europe + 328Cao K.I. et al. [31]20172001–2014Europe −/+ 101.Ioannides P. J. et al. [32]20152004–2014USA − 134Szymańska A. et al. [33]20151983–2013Europe −/+ 846Rougier G. et al. [34]20232011–2021Europe + 399Browne J.D. et al. [7]19931981–1990Europe − 159Zanoletti E.; Mazzoni A. [35]20061990–2003Europe + 1616Netterville J.L. et al. [36]19981978–1998USA + 4646Bradshaw J.W.; Jansen J.C. [37]20051974–2003Europe + 4848Miller R.B. et al. [38]20001977–1996USA + 1919Lozano F.S. et al. [5]20081991–2006Europe − 426Eriksen C. et al. [39]1991-USA − 98González-Orús A.R. et al. [40]20151990–2013Europe + 12217Lack E.E. et al. [41]19771937–1975USA − 6913Urquhart A.C. et al.* [12]19941972–1992USA − 1919Kollert M. et al. [1]20061981–2004Europe − 2213Shin S–H. et al. [42]20121989–2009Europe + 2222Rana M.U. et al. [2]20211990–2017Europe − 634Ding Y. et al. [4]20212015–2020Asia −/+ 62Smith J.D. et al. [43]20172000–2017USA + 19421Erickson D. et al. [44]20011978–1998USA − 20426Merzouqi B. et al.** [15]20212010–2020Africa − 265 Description of the studies included. Quality assessment: + (good ≥ 80%), −/+ fair (80–50%), − poor (≤ 50%). *Multicentric, **prospective The 41 papers mainly reported miscellaneous cases of PGs on a total of 2169 patients, with 501 fitting our criteria for inclusion. Excluding six studies that focused exclusively on VPGs, the enrollment rate in the remaining studies was 22.8%. Furthermore, although almost 80% of the articles reported an enrollment period ≥ 10 years, 43.9% of the studies included fewer than 10 cases of VPGs. More than 90% of the studies were published in Europe or North America (22 and 15 articles, respectively), with only 3 articles in Asia and 1 in Africa. This probably reflects a publication bias rather than a true difference in the incidence of VPGs. We observed a growing interest in this pathology, as shown by the number of publications over the last 25 years. Despite the primacy of surgery in the absolute number of studies, we recorded a greater proportional increase in articles regarding radiotherapy from 2010 to date (Fig. 2). Temporal distribution of studies. s = surgery; rt = radiotherapy; mix = surgery and/or radiotherapy and/or wait and see Clinical data and treatment details The mean age of patients included was 44.7 years (M 37.74%, F 62.26%); paragangliomas were mostly primary (91.99%), and benign (96.01%). Multiple tumors were found in 25.15% of cases, while a hereditary basis was present in 15.17%. Table 2 summarizes the patient characteristics, treatments and outcomes. Table 2Clinical characteristics, treatment and outcomesN. of patients (%)Gender F198 (62.26%) M120 (37.74%)Mean age44.70 years (18–85 years)Ereditary condition76 (15.17%)Multiple paragangliomas131 (25.15%)Malignant VP20 (3.99%)Primary VP (337 pts)310 (91.99%)Recurrent VP (337 pts)27 (8.01%)Treatment Wait and see66 (13.17%) Surgery325 (64.87%) Surgery + RT9 (1.80%) RT101 (20.16%)Surgical outcomes Gross total resection241 (74.15%) Subtotal resection20 (6.15%) Unknown64 (19.69%)RT outcomes Complete/partial response42 (41.58%) No response2 (1.98%) Unknown59 (58.41%)Mean follow-up (389 pts)69.32 monthsLocal recurrence14 (3.60%)Death for disease2 (0.51%)Death for unrelated causes5 (1.28%) Clinical characteristics, treatment and outcomes A total of 325 patients (64.87%) underwent surgery, 101 (20.16%) RT, and 9 (1.80%) both treatments; a wait-and-see policy was preferred for the remaining 66 (13.17%). Embolization was mainly performed before surgery or radiotherapy (93 patients); only one patient was treated with palliative embolization alone, with no progression recorded. Concerning surgical outcomes, a gross total resection was achieved in 74.15% of cases, a subtotal excision in 6.15%, while the extent of surgery was unknown in the remaining 19.69%. Unfortunately, an assessment of the efficacy of RT was not specified in most cases (58.41%) and there was no consensus in defining response to radiation treatment. In fact, most studies generally considered RT successful if no clinical or radiological progression was recorded during follow-up; only 2 articles [21, 34] classified response according to RECIST criteria [45, 46], whereas Tamaki et al. [9] considered a cut-off of 15% in increase of volume to evaluate the outcomes of treatment. Nevertheless, an absent response/progression was described only in two cases (1.98%), while a complete or partial response was reported in 41.58% of patients. After a mean follow-up of 69 months, there were 14 local recurrences (3.60%), only one after RT. Complications and cranial nerve resection Publications describing surgical experience were unfortunately heterogeneous and surgical complications, such as nerve deficit, were not methodically presented. Only 18 of 41 publications evaluated pre- and postoperative function of cranial nerves. A vagus nerve deficit was reported at presentation in 74 patients (38.74%). As expected, cranial nerve X was unavoidably sacrificed during surgery, resulting in dysphonia/vocal cord palsy, even if a real prevalence could not be assessed. The articles by Kunzel et al. and Miller et al. were the only ones explicitly stating the preservation of vocal cord motility following partial excision [30, 38]; similarly, Browne et al. described a case of functional rehabilitation after intraoperative nerve grafting [7]. Concerning other cranial nerves, concomitant hypoglossal nerve palsy at presentation was the most frequently observed (20 patients, 10,5% of cases), followed by glossopharyngeal and accessory nerve (17 and 15 patients, respectively). Excluding vagus nerve sacrifice, 158 cases of postoperative paralyses were reported, which is 2.6-times higher than the preoperative situation. Cranial nerves VII and XII showed the greatest increase (3.0- and 3.2-fold, respectively). Table 3 reports cranial nerve impairments before and after surgery. Table 3Cranial nerve impairments in patients undergoing surgeryAuthorsX deficit at presentationAdditional cranial nerve deficit at presentationX function preservationPostop cranial nerve deficitVIIIXXIXIISympateticVIIIXXIXIISympateticLi, L. et al. [16]000000000000Guha, A. et al. [13]000010000010Papaspyrou K. et al. [20]412100235471Michelozzi C. et al. [23]400010201020Persky MS. et al. [10]101100111110Paris J. et al. [24]200001200011Künzel J. et al. [30]100000102120Szymańska A. et al. [33]433440233440Browne JD et al. [7]411110113230Zanoletti E et al. [35]503320023220Bradshaw JW et al. [37]200000002151Miller RB. Et al. [38]700042200092Eriksen C. et al. [39]100000002140González O. et al. [40]1000000033241Urquhart AC et al. [12]900000000051Kollert M et al. [1]301120028350Shin S–H et al. [42]1006450008670Smith JD et al. [43]700000002022Total (191 patients)745171520313154327649 Cranial nerve impairments in patients undergoing surgery As an expected outcome of surgery, mild dysphagia may have been implied in many cases and only moderate to severe forms may have been reported; nonetheless, it was by far the most common surgical complication (48 cases of the 204 reporting complications, 23.53%), although only 14 patients underwent percutaneous endoscopic gastrostomy to guarantee nutrition. All other surgical complications were less frequent, as highlighted in Table 4, but included extremely severe events, as internal carotid artery rupture (5 cases). In this regard, 2 of the 5 events happened during surgery for multiple paragangliomas, suggesting a carotid body tumor may have been simultaneously treated in these cases. Table 4Surgical complicationsICA rupture5 (2.45%)Wound infection8 (3.92%)CFS fistula4 (1.96%)Aspiration/dysphagia48 (23.53%)Stroke5 (2.45%)Bleeding3 (1.47%)Pneumothorax2 (0.98%)Endocarditis1 (0.49%) Surgical complications Functional outcomes after radiotherapy, such as acute and late toxicity, were rarely reported; nevertheless, when provided, toxicities were mainly mild (see Table 5 for specific complications in VPGs), and additional cranial nerve palsies after treatment were almost never observed. Mendenhall et al. [21] stated that cranial nerve palsies resolved in 6 of 36 patients and no patient developed a new cranial nerve palsy after RT; similar results were reported by Michelozzi et al. [23] and Ioannides et al. [32]. Table 5Toxicities after radiation therapy (only studies that isolated complications for VPGs are included)Authors (Year)Treatment type (N. of pts)Severity of complicationsType of complications (% of pts)Rougier et al. (2022) [33]IMRT (9 VPGs considered)33% G I; 11% G II; 0% G III, G IV and G VG I dermatitis, G I nausea (33% each); G II alopecia, G II nausea (11% each)Künzel et al. (2013) [31]SRT (2 VPGs considered)NA0% of cranial nerves deficitAbbreviations: pts patients, G grade, IMRT intensity-modulated radiotherapy, SRT stereotactic radiotherapy, NA not available Toxicities after radiation therapy (only studies that isolated complications for VPGs are included) Abbreviations: pts patients, G grade, IMRT intensity-modulated radiotherapy, SRT stereotactic radiotherapy, NA not available Discussion PGs represent less than 1% of all head and neck neoplasms and this uncommon occurrence accounts for the fragmentary evidence on their natural history and appropriate treatment, with sporadic reports or case series with few patients, often with heterogeneous baseline conditions or treatments [23, 47]. Although HNPGs are often reported collectively, VPGs have peculiar features that should be considered during management, and the present systematic review was designed to provide specific evidence on VPGs, when feasible, while data retrieved from other HNPGs will be necessary otherwise [4, 47]. We observed a vast majority of monocentric and retrospective studies; furthermore, VPGs were often outnumbered by other HNPGs, as confirmed by the low enrollment rate (around 20%). In this panorama, information regarding VPGs was incomplete and difficult to deduce. The abovementioned uncertainty on HNPGs firstly applies to the growth rate: PGs are considered slow-growing tumors, but with unpredictable behavior [9, 48, 49]. In this regard, Tamaki et al. [9] reported 17 of 29 observed HNPGs having a > 15% growth (5 of 12 VPGs) and only 2 with worsening symptoms [17]; Jansen et al. similarly described 70 growing tumors out of 157 HNPGs (12 of 29 VPGs, median growth rate 1.6 mm/year), with 17 tumors (6 growing) determining complications [48]. Interestingly, being growing tumors did not relate to complications, as already found in a previous report [37, 48]. Young age (especially < 50 years), familial cases, and intermediate size have been linked to faster increase in size [48]. Analogously, malignancy in PGs and pheochromocytomas has been heterogeneously considered, although most authors identify it with the presence of distant metastasis in non-chromaffin tissues (to avoid misinterpretation of multifocal disease), and not local infiltration [3, 47, 50, 51]. This has a profound clinical impact, as there is still no definitive method (even histological) to identify a malignant disease except for metastasis. As malignancy may be detected after primary treatment, its incidence may depend on the duration of follow-up [50, 51]. VPGs have been considered more frequently malignant than most other HNPGs, with an estimated prevalence of 10% in a previous report [50]; our systematic review seems to downsize this increase in risk, with a prevalence close to 4% [3, 50]. While surgery has been historically considered the treatment of choice for HNPGs, the last few decades have seen an increasing interest in other approaches, mostly wait-and-scan and RT [3, 9, 37, 48, 49, 52]. A non-interventional strategy is sustained by the abovementioned evidence of slow growth, low rate of malignancy, and low rate of complications/local invasiveness, especially compared to surgery. On the other hand, evidence from the same studies may support treatment for different reasons: (1) the follow-up is not particularly long, suggesting malignancy or growth in size may present later; the study by Jansen et al. ultimately found 90% of HNPGs growing within 52 months [48]; (2) patients chosen for a wait-and-see policy may have suffered a selection bias (e.g., less aggressive PGs) compared to treated lesions in some studies; (3) studies from the Netherlands analyzed patients living at low altitudes, while low oxygen pressure has been associated with an increase in growth of PGs, limiting the generalizability of the results; (4) there is a lack of predictors of complications, as local invasiveness (e.g., neural deficits) often derives from patients with stable tumors [9, 37, 48, 52]. For these reasons, a wait-and-scan approach has been proposed, especially for elderly patients (having a lower life expectancy for development of complications), unifocal and small VPGs [9, 48, 52]. When an observational approach is not chosen, the main options are surgery (with eventual presurgical embolization) and radiotherapy/stereotactic radiosurgery (SRS). Surgery has been for long the preferred option but, as highlighted by our systematic review, it is burdened with frequent, disabling consequences. Vagal nerve palsy appears almost unavoidable: in our review the preservation of vagal nerve function was specifically stated for only three patients. Cranial nerve VII, IX, XI, and XII and cervical sympathetic chain palsies were also found in 7.9%, 21.5%, 14.1%, 33.5% and 4.7% of cases, respectively; aspiration and dysphagia were consequently the most common symptoms, although only 14 patients needed a percutaneous gastrostomy. Considering all cranial nerves, every surgical procedure on VPGs is burdened by a 2.6-fold increase in neural deficits, which is consistent with previous reports. This incidence is hardly compatible with the excision of a benign, slow-growing neoplasm, and this is the reason behind a shift towards RT/SRS in the last decades, as confirmed by an increase in the frequency of articles on RT after 2000 in the present review. The main advantage of surgery is the possibility of a complete eradication of the disease, which makes it a preferred option by most authors for malignant and secreting tumors (respectively 3.99% and 0.6% in our review). Surgery is also particularly valid when cranial nerve X has already been affected, although post-surgical worsening may still result due to a partial remaining functionality (for example in laryngeal sensitivity). A concurrent dissection of the closest lymph nodes has been suggested to ensure better surgical access or to rule out metastasis in malignant cases and/or other undiagnosed smaller PGs, especially in familial cases [50]. A pre-surgical embolization has been widely performed to reduce vascular flow in HNPGs, mostly for non-carotid body tumors, and a recent systematic review reported technical success (defined as an 80% reduction in blood flow) in 79% of cases, with complications in 3.8% of patients [53]. Some authors have reported subtotal resection to preserve the function of the cranial nerve X, thereby possibly avoiding the most common consequence of the procedure on the quality of life; on the other hand, leaving neoplastic tissue in the field may lessen the rate of control and only partially reduce the mass effect. The number of cases of subtotal resection was too low and the follow-up too short to confirm or deny these assumptions and there is no evidence on the extent of surgical excision that can preserve nerve function; thus, in our opinion, partial surgery should be considered in the context of a more modern, eventually multimodal approach to VPGs. More studies are needed to evaluate the potential applicability of recurrent debulking surgery (in consideration of the low rate of growth), or as a pre-RT procedure to reduce mass dimension in voluminous PGs (considering the higher response rates of smaller tumors to RT). As shown (Fig. 2), RT has had increasing relevance in the treatment of VPGs in the last decades. Unfortunately, reports on RT of VPGs are mostly lacking exhaustive details on outcomes, complications, and follow-up: in 56.44% of cases the result of RT on tumor size is, in fact, not specified. Nonetheless, an absent response was incredibly rare (1.98%). It is worth stressing that RT does not lead to a mass dissolution, and the response is therefore defined by lesion stability over time, tumor shrinkage, or lack of macroscopic progression of disease. Our finding is consistent with a previous systematic review by Fatima et al. selectively studying the efficacy of stereotactic radiosurgery on HNPGs, that found, among patients with radiographic outcome control, stability, shrinkage, or progression of the mass in 48.6%, 45.6%, and 5.8% of cases, respectively; the same publication reported a higher control rate for smaller compared to larger tumors [54]. Analogously, the presence or absence of cranial nerve palsies is not specified in most reports, although it seems to generally imply the persistence of nerve functionality. Only three articles reported cranial status after RT; the study with the largest number of cases reported functional recovery of the vagus nerve in 16.7% of cases (6 of 36), with no additional nerve deficits [21]. The abovementioned systematic review on HNPGs by Fatima et al. found improvement, stability, or worsening of neural status in 48.7%, 39.3%, and 1.2% of patients treated with SRS, respectively [54]. In this context, a comparison with complete surgical excision appears difficult, because many studies have limited follow-up time, and dimensional stability over years following RT (which can still represent the natural history of the disease) does not undoubtedly rule out a future regrowth; in addition, many biases (such as the choice of surgery for larger tumors for the fear of ineffective RT) may apply. On the other hand, higher levels of evidence or comparative studies are limited by the low prevalence of disease, and the marked differences in complications seem to unequivocally support RT over complete surgical excision at least for cases of small, non-secreting tumors with preserved nerve functionality. Recent years have seen a progressive increase in the indications of proton therapy for different neoplasms; its main advantage over photon radiotherapy is the significant reduction of radiation dose to surrounding tissues, due to a higher concentration at the end of the beam range and a virtually absent exit dose [30, 31, 55, 56]. These features, which may apply to all patients and neoplasms, are particularly valuable for subjects affected by VPGs, who are usually young and with a survival expectancy of various decades. Its main limit is the confined availability, which invariably affects the pool of studies in the literature, with only three cohorts of patients (64 patients of HNPGs in total, 8 VPGs) in our systematic review (all articles reporting mixed HNPGs and not only VPGs). On the other hand, these studies have a long median follow-up time, and the complications and efficacy are methodically displayed [31, 32, 55, 56]. In particular, dimensional control or shrinkage was almost univocal, with only two cases of progression during follow-up; analogously, acute and late toxicities were mostly mild (grade 1 or 2), with some cases of improvement in cranial nerve deficits [31, 32, 55, 56]. The study by Ioannides et al. also found a faster volumetric shrinkage compared to conventional RT, although some bias may apply [32]. In general, due to the limited number of cases, a comparison between the two techniques is still not possible, and the supposed advantages of proton therapy are still based on the theoretical mechanism of action. A feature which is worth stressing is that the reported complications, especially for a mainly benign neoplasm, should not be considered per se, but weighted by their impact on the patient’s quality of life (QoL). Unfortunately, the use of self-reported questionnaires is quite limited and heterogeneous: some studies [57] adopted them to prove the significant burden of HNPGs on the QoL, but their use in the evaluation of different treatments is reported only in a few papers. Between them. Cao et al. [31], administered the European Organization for Research and Treatment of Cancer QoL questionnaire (EORTC QLQ-C30) to their proton-therapy treated HNPGs patients only after therapy, reporting a 66.5/100 mean value (where 100 is the perfect QoL) [58]; used EORTC QLQ-C30 and EORTC H&N35 to find a statistically significant worsening in QoL when a multimodality treatment was used to treat HNPGs when compared to RT alone; Canu et al. [59] compared a surgical and a wait-and-see strategy using a variety of questionnaires, finding conflicting results. The main strength of the present article lays in its strict methodology for a comprehensive literature review; still, registration on PROSPERO would have implemented its transparency and reproducibility. The exclusion of papers reporting less than 5 cases aimed to reduce the possibility of including variables related to insufficient expertise on the disease, but the limited number of reports in the literature may have led to a lower recognition of rare complications and hindered the generalizability of the results. The main limitations are intrinsically related to the limited and heterogeneous reports in the literature, as suggested by the poor median quality of the papers included in the review, which precluded a meta-analysis and a more structured comparison of outcomes and complications. For future papers we suggest a systematic description of the clinical features of the patient (including comorbidities and previous other PGs), the neoplasm (size, site, evolutions over time, infiltrations of surrounding structures, neural deficits at presentation, impact on the QoL), treatment’s motivated choice (wait and see with eventual modifications, surgical extent/radicality, type and intensity of RT), outcome over time, and complications (including a systematic evaluation of neural deficits, quality of life and need of subsequent surgery for nerve impairment). Additionally, vagal paragangliomas are outnumbered by other head and neck PGs, but they typically carry different surgical risks and nerve complications, so they are best debated separately. A further limitation of the present review lies in the geographical distribution of the included studies, with the predominance of them (more than 90%) conducted in Europe and North America. This may reflect not only differences in research activity but also a potential publication bias, whereby studies from other regions may be underrepresented or unpublished in indexed journals. This geographical concentration may introduce a form of contextual bias, as the findings may reflect region-specific healthcare systems, cultural factors, or research priorities. Consequently, the generalizability of our findings to healthcare systems or patient populations outside of Western countries may be limited. This imbalance should be taken into account when interpreting the current evidence base and planning future multicentric, globally representative studies. Conclusions and future perspectives At present, there are many treatment options for head and neck VPGs, but surgery and radiotherapy/SRS still represent the cornerstone of the therapeutic approach. Surgery allows removal of the disease and is almost invariably preferred for secerning and malignant VPGs; however, vagal nerve impairment is often unavoidable, implying vocal cord paralysis and sometimes dysphonia and dysphagia. Data in the literature are still lacking for a methodical comparison of the rate of success and complications, but RT/SRS seems favorable, at least for small, non-secreting tumors with preserved nerve functionality, due to the apparently lower frequency of complications. A tailored approach considering the patient’s performance status, extension of tumor and personal expectations is nevertheless always advised, especially with PGs having features different from the abovementioned. To overcome the surgical main drawback, partial or nerve-sparing resection has been proposed, but functional preservation of cranial nerve X is rarely achieved [30, 38]. Recently, Kampel et al. described intracapsular resection of cervical vagal schwannomas with intraoperative nerve monitoring: vagal nerve palsy occurred in one of 4 patients, although one patient presented residual disease after piecemeal resection [60]. As schwannomas and PGs both present a thin fibrous capsule, this technique may apply to VPGs. Nevertheless, targeted clinical studies are needed to confirm this assumption and confirm the functional preservation of the nerve. Another interesting development regards the treatment of unresectable or metastatic disease. Ongoing clinical trials are evaluating the use of theranostic: a radioactively labeled drug (177Lutetium DOTATATE) binds specifically to SSTR, expressed by 68 Ga-DOTATATE positive paraganglioma, delivering radiation. Despite the restrictive indication, Satapathy et al. [61] detected a response rate and disease control rate of 25% and 84.5%, respectively. Treatment-related side effects are rare and mainly encompass lymphopenia, thrombocytopenia, and renal dysfunction. However, data on the late effects of the treatment, including clinical outcomes, are still lacking. Supplementary Information Supplementary Material 1. Supplementary Material 2. Acknowledgements We thank the institutional board of FPO-IRCC Candiolo Cancer Institute. Abbreviations Vagal paraganglioma/vagal paragangliomas Paraganglioma/paragangliomas Head and neck paraganglioma(s) Carotid body tumor(s) Radiotherapy Stereotactic radiosurgery Preferred Reporting Items for Systematic Reviews and Meta-Analyses Response Evaluation Criteria in Solid Tumors National Heart, Lung, and Blood Institute Study Quality Assessment Tools Somatostatin receptor Succinate dehydrogenase subunit genes (SDHA, SDHB, SDHC, SDHD, SDHAF2) Quality of life European Organization for Research and Treatment of Cancer QoL questionnaire European Organisation for Research and Treatment of Cancer, Head and Neck module, 35 items Authors’ contributions Conceptualization, A.R., S.B., and L.V.C.; investigation, E.S., L.V.C., A.F., D.D.S., and P.L.V.; data curation, E.S., L.V.C., A.F., D.D.S., and P.L.V.; writing—original draft preparation, A.R., L.V.C., and S.B.; writing—review and editing, L.V.C., A.R., and S.B.; project administration, S.B.; visualization, L.V.C. and S.B.; supervision, L.V.C. and S.B.. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by the project FPRC 5xmille Ministero Salute 2018 ADVANCE and Italian Ministry of Health, Ricerca Corrente 2025. Data availability No new data were created in this study. All the papers considered for the review are publicly available in PubMed, Embase or Scopus. Declarations Ethics approval and consent to participate No statement of the institutional review board is required to perform a systematic review and meta-analysis. Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Data Availability Statement No new data were created in this study. All the papers considered for the review are publicly available in PubMed, Embase or Scopus.