Tumoral parkinsonism—Parkinsonism secondary to brain tumors, paraneoplastic syndromes, intracranial malformations, or oncological intervention, and the effect of dopaminergic treatment

Abstract Introduction Secondary tumoral parkinsonism is a rare phenomenon that develops as a direct or indirect result of brain neoplasms or related conditions. Objectives The first objective was to explore to what extent brain neoplasms, cavernomas, cysts, paraneoplastic syndromes (PNSs), and oncological treatment methods cause parkinsonism. The second objective was to investigate the effect of dopaminergic therapy on the symptomatology in patients with tumoral parkinsonism. Methods A systematic literature review was conducted in the databases PubMed and Embase. Search terms like “secondary parkinsonism,” “astrocytoma,” and “cranial irradiation” were used. Articles fulfilling inclusion criteria were included in the review. Results Out of 316 identified articles from the defined database search strategies, 56 were included in the detailed review. The studies, which were mostly case reports, provided research concerning tumoral parkinsonism and related conditions. It was found that various types of primary brain tumors, such as astrocytoma and meningioma, and more seldom brain metastases, can cause tumoral parkinsonism. Parkinsonism secondary to PNSs, cavernomas, cysts, as well as oncological treatments was reported. Twenty‐five of the 56 included studies had tried initiating dopaminergic therapy, and of these 44% reported no, 48% low to moderate, and 8% excellent effect on motor symptomatology. Conclusion Brain neoplasms, PNSs, certain intracranial malformations, and oncological treatments can cause parkinsonism. Dopaminergic therapy has relatively benign side effects and may relieve motor and nonmotor symptomatology in patients with tumoral parkinsonism. Dopaminergic therapy, particularly levodopa, should therefore be considered in patients with tumoral parkinsonism.


INTRODUCTION
Secondary parkinsonism caused by tumors, or in short tumoral parkinsonism, is defined as parkinsonism developing as a direct or indirect result of tumors, most frequently through mesencephalic infiltration or compression (Bhatoe, 1999). However, parkinsonian symptomatology in patients with brain tumors is rarely described. In a prospective evaluation, eight out of 907 patients with supratentorial brain tumors were found to suffer from some type of secondary parkinsonism-indicating a prevalence of 0.3% (Krauss et al., 1995). The most common localization of these neoplasms is supratentorial. Tumor entities are most often meningioma, followed by glioma (usually astrocytoma), central nervous system (CNS) lymphoma, and craniopharyngioma (Höllerhage, 2019). Various types of brain cysts, cavernomas, and brain metastases have also been reported to cause secondary parkinsonism (Chang et al., 2000;Hortelano et al., 2010;Ishihara et al., 2011). Other relevant medical causes in the context of a broader definition of tumoral parkinsonism include paraneoplastic syndromes (PNSs) (Golbe et al., 1989;Topcular et al., 2013) and iatrogenic oncological parkinsonism (Franchino et al., 2019;Markman et al., 1985;Skiming et al., 2003;Wenning et al., 1999). Surgical removal or complete resection of the tumor generally has a good effect on parkinsonian symptomatology, often resulting in marked improvement or complete reversal of symptoms. As for dopaminergic therapy, the general opinion is that it has little to no effect, but different articles have reported conflicting results. No controlled trials have been performed (Höllerhage, 2019).
This systematic synthesis of the literature was motivated by the clinical experience of six patients suffering from parkinsonism secondary to different types of brain neoplasms (Table 1). Interestingly, all these cases improved significantly after initiation of dopaminergic therapy.
These findings contrast the current oncological and neurological consensus (Höllerhage, 2019). As there were no current compilations of data on this topic, this justified a crude analysis of outcome of dopaminergic therapy on tumoral parkinsonism patients, primarily col- The aim was to make this review useful in a practical clinical setting, by categorizing relevant data according to tumoral entity, to facilitate finding information about a specific tumor entity or phenomenon of interest.

Search strategy
The search for articles in this review was performed through the medical databases PubMed and Embase. Two separate searches were carried out in PubMed and one in Embase. Examples of terms included in these searches are "secondary parkinsonism," "astrocytoma," and "cranial irradiation." Synonyms and terms related to all search terms were allowed and used. The search strategy required the terms to be found in either the title or abstract of the articles, or to be subject headings such as MeSH terms or Emtree terms of the article. For a complete presentation of the database search, specifying all strategies and terms utilized in this literature review, please see Table 2.

Abstract inspection
The articles collected from the database search were inspected by reading the title, abstract, and keywords.
Full-text reading was carried out on articles fulfilling the required inclusion criteria: • The study had to focus on a subtype or aspect of parkinsonism secondary to ○ Brain tumors OR ○ PNSs OR ○ Nontumoral intracranial malformations OR ○ Oncological intervention. • The language of the article must be English.
• Appropriate ethical considerations regarding informed consent, risk of harm, confidentiality and anonymity, and conflict of interest.
The exclusion criteria were as follows: original research must be presented, and review articles were excluded. TA B L E 1 Patients with parkinsonism secondary to brain neoplasms that served as the primary motivation for conducting the review. Note: Two separate searches were carried out in PubMed and one in Embase. Articles identified with this systematic search approach were screened for eligibility before inclusion in the study. Abbreviations: exp, exploded-Emtree term; T/A, Title/Abstract; ti,ab,kw, Title, Abstract, Author Keywords.

Full-text reading and detailed review
Articles aligning with abovementioned criteria were read in full. New articles discovered in the reference list of other articles or by specific searches could also be further investigated and added to full-text review. However, these types of articles were reported as a separate category in the flow diagram presented in Figure 1.

Meningioma
Based on the number of case reports in the literature, the most common type of CNS tumor to be reported to cause parkinsonism is meningioma (Höllerhage, 2019), also being the second most common type of tumor in the CNS. However, due to insufficient epidemiologic data, it is impossible to fairly assess the exact incidence of menin-  Although the other studies did not present such thorough imaging workups, this study might suggest that parkinsonism can be the result of either direct pressure from the tumor or a metabolic dysfunction.

Glioma
The second most common type of brain tumor to be reported to cause parkinsonism is glioma, predominantly of the subtype astrocytoma (Höllerhage, 2019). Ho et al. reviewed the medical history of a 60-year-old man who developed right-sided hemiparkinsonism, most notably resting tremor and bradykinesia. Initially, the patient was thought to suffer from idiopathic Parkinson's and was consequently treated with levodopa. Since this had no effect, MRI was conducted for further investigation, revealing an infiltrative astrocytoma span-ning from the left mesial temporal lobe to the left basal ganglion and insula. Reversal of initial symptoms was observed after resection, but an ipsilateral hemiparesis presented postoperatively. The histopathological diagnosis was high-grade astrocytoma (Ho et al., 2008). Others have reported similar symptomatology as a result of corpus callosum astrocytoma (Arcaya Navarro et al., 1986) and brain stem cystic astrocytoma (Cicarelli et al., 1999 Navarro et al., 1986;Choi et al., 2012;Cicarelli et al., 1999;Ho et al., 2008;Wächter et al., 2011).

CNS lymphoma
Primary CNS lymphoma is a rare disease, primarily affecting older people, and secondary parkinsonism due to CNS lymphoma is extraor-

Other types of primary brain tumors
Theoretically, any brain tumor can cause secondary parkinsonism as a result of distant mass effect and metabolic and vascular disturbances (Bhatoe, 1999). Dolendo et al. reported a 14-year-old male with immature teratoma in the pineal gland exhibiting parkinsonism. A ventriculoperitoneal shunt improved the hydrocephalus, but dopaminergic therapy did not improve parkinsonian symptoms. Symptoms did improve after reprogramming the shunt. Despite numerous excisions and chemotherapy, the tumor continued to enlarge, and the patient was assessed to suffer from parkinsonism secondary to growing teratoma syndrome (Dolendo et al., 2003). Biochemical studies on a patient with craniopharyngioma similarly suggested the atypical tumor as the cause of parkinsonism (García de Y'ebenes et al., 1982). Parkinsonism in children has been described in two cases of mesencephalic tumors (Pohle & Krauss, 1999

Brain metastasis
Brain metastases can originate from a wide range of neoplasms, but most derive from melanoma, lung, breast, colon, or kidney cancer. CT (Chang et al., 2000). There is a wide agreement that patients with extracranial cancer who rapidly develop parkinsonian symptoms must be investigated for brain metastases in order to facilitate adequate therapy and improve disease prognosis (Chang et al., 2000;Hortelano et al., 2010;Ishihara et al., 2011 However, remission was not observed when trying to taper off medications, and the patient thus had to continue with initial therapy indefinitely. There was no need to increase dosages (Pereira et al., 2013). Other examples of parkinsonism emerging during chemotherapy include a patient who was also administered highdose metoclopramide for antiemetic purposes during the treatment (Markman et al., 1985), a dialysis-supported patient who received high-dose unspecified chemotherapy for multiple myeloma (Fleming & Mangino, 1997), and a patient who was treated with unspecified chemotherapy for non-Hodgkin lymphoma (Howell & Sagar, 1994). Metoclopramide is a well-known cause to drug-induced parkinsonism.

Oncological parkinsonism in children
It is extremely rare for idiopathic PD to develop in children, and this type, which is referred to as juvenile parkinsonism, is often associated with specific, high-risk genes. Secondary parkinsonian symptoms in children following cancer treatment, although still uncommon, are often possible to prevent or reverse by optimizing or altering treatment. Furthermore, secondary parkinsonism after cancer therapy has been reported to be more common in general pediatric hospitals than rare genetic parkinsonian movement disorders (Pranzatelli et al., 1994). Cases of secondary parkinsonism in children have been described after craniospinal radiotherapy in both teenagers (Bernard & Chouinard, 2011;Voermans et al., 2006) and infants (Skiming et al., 2003), with somewhat varying responses to dopaminergic therapy. Other cases of secondary parkinsonism include an infant with acute leukemia who was treated with the chemotherapeutics vincristine and adriamycin (Boranic & Raci, 1979) and three children with refractory leukemia who were treated with bone marrow transplantation and high-dose amphotericin B (Mott et al., 1995). Pranzatelli et al. reported the largest study so far on children with a spectrum of secondary parkinsonism, including six hospitalized children with parkinsonism. The symptoms of some of these patients were interpreted to be the result of cancer treatment, but the study contained a wide array of suggested etiologies in the different cases.
All children improved dramatically, most having complete reversal of symptoms and none having to continue dopaminergic therapy (Pranzatelli et al., 1994).
In the case of Kleib et al., the patient had been taking levodopa for 4 months without effect, before MRI revealed a meningioma and surgery was performed (Kleib et al., 2016). Ho et al. prescribed lev-odopa and dopamine agonists, believing that the patient suffered from idiopathic PD, but due to the unresponsiveness of medication, brain CT was conducted and showed an astrocytoma (Ho et al., 2008). Modreanu et al. described a long process of levodopa adjustments in a patient with a known cavernoma. Initially, the patient was medicated with low-dose levodopa, but the doses were gradually increased to 750 mg per day. However, the patient still claimed a lack of response, and this was later confirmed with negative levodopa and apomorphine tests. Levodopa treatment was discontinued, and no worsening of motor or nonmotor symptoms was observed (Modreanu et al., 2020). Alp et al. and Ertan et al. both reported patients with cavernomas, who refused surgical intervention and did not respond to levodopa even at high doses (Alp et al., 2009;Ertan et al., 2005). The growing teratoma syndrome with subsequent parkinsonism investigated by Dolendo et al. was managed with levodopa-carbidopa and amantadine, but to no avail (Dolendo et al., 2003). Levodopa was also unhelpful in the CNS lymphoma patient examined by Ishihara et al. (2011). Other cases where high doses of levodopa-carbidopa were ineffective in relieving parkinsonian symptoms include two patients with glioblastoma (Ruiz-Escribano Menchen et al., 2020) and a teenager developing parkinsonism postradiotherapy (Bernard & Chouinard, 2011). Li discussed an unusual case of parkinsonian symptoms caused by breast ductal adenocarcinoma, who received levodopa to ease neurological deficits, but with no effect (Li, 2019).
Yet another breast ductal adenocarcinoma leading to parkinsonian symptoms was illuminated by Golbe et al. This patient was treated with several drugs, including anticholinergics, baclofen, diazepam, levodopa-carbidopa, and plasmapheresis, all resulting in neither positive nor negative change in clinical presentation (Golbe et al., 1989).
No studies included in this review have reported negative or unexpected side effects of dopaminergic treatment in patients suffering from tumoral parkinsonism.

Low to moderate effect
Twelve of the 25 (48%) studies initiating dopaminergic therapy in individual patients suffering from some type of tumoral parkinsonism found partial effect on clinical symptomatology (Artusi et al., 2015;de Sèze et al., 1998;Dunbar et al., 2012;Franchino et al., 2019;Gherardi et al., 1985;Ho et al., 2019;Malomo & Emejulu, 2008;Pranzatelli et al., 1994;Reddy et al., 2022;Voermans et al., 2006;Wächter et al., 2011;Wimmer et al., 2020). Partial effect in this context is defined as low to moderate reversal of parkinsonians symptoms as a consequence of therapy with some sort of dopaminergic agent. Wimmer et al. wrote about a patient who was treated with levodopa and pramipexole for hemiparkinsonism associated with an arachnoid cyst. Although the treatment was effective at first, symptoms worsened over time as the underlying cysts progressed, thus leading to low treatment effect (Wimmer et al., 2020 (Gherardi et al., 1985). In both cases, the symptoms gradually worsened, and this motivated neuroimaging and subsequent findings of brain neoplasms. This chain of initial low to moderate effect of dopaminergic therapy in what is thought to be idiopathic PD, followed by unusually rapid exacerbation in clinical presentation, and finally neuroimaging and the discovery of an unknown brain tumor, is a recurring pattern in several tumoral parkinsonism case reports (de Sèze et al., 1998;Gherardi et al., 1985;Wächter et al., 2011). Several studies have described low to moderate effect of dopaminergic therapy in parkinsonism related to iatrogenic effects of tumor treatment following surgery (Malomo & Emejulu, 2008) and radiotherapy (Artusi et al., 2015;Franchino et al., 2019;Reddy et al., 2022;Voermans et al., 2006). Dunbar et al. investigated the potential use of basal ganglia drugs as palliative medication in a cohort of 21 patients with latestage brain tumors. They found that agents such as methylphenidate, modafinil, levodopa, and amantadine relieve at least one major parkinsonian symptom in 86% of the included patients, thus indicating that these drugs could have utility in palliative care of brain cancer patients with parkinsonism (Dunbar et al., 2012).

Excellent effect
Two of the 25 (8%) studies initiating dopaminergic therapy on patients suffering from some type of tumoral parkinsonism found excellent effect on clinical symptomatology (Pereira et al., 2013;Straube & Sigel, 1988). Excellent effect in this context is defined as complete reversal of parkinsonian symptoms as a consequence of therapy with some sort of dopaminergic agent. Straube et al. treated a low-grade glioma patient, presenting with resting tremor, rigidity, akinesia, and slight right leg paresis, with 400 mg of levodopa-carbidopa and 12.5 mg of bromocriptine per day. Complete reversal of symptoms was observed immediately after treatment initiation, and there had been no deterioration 5 years after onset, without altering dosages. The slight right leg paresis did not regress nor progress during this time (Straube & Sigel, 1988). Pereira et al. investigated three middle-aged female patients who developed parkinsonism secondary to oncological treatment, more specifically systematic chemotherapy for lymphoma, whole-brain radiotherapy for glioma, and motor cortex radiosurgery for pulmonary adenocarcinoma metastasis, respectively. Parkinsonian symptoms started after 1 year of oncological intervention and consisted of mild bradykinesia and rigidity in all cases. Additionally, one patient suffered from moderate tremor. Anticholinergic and dopaminergic therapy had excellent results in all cases, and none of the patients had to increase dopaminergic treatment 2-5 years after onset.
However, none of the patients had reversal of symptoms without medication and required persistent antiparkinsonian therapy (Pereira et al., 2013). Both these studies described full symptomatic reversal with dopaminergic therapy, but insufficient clinical presentation without treatment.

DISCUSSION
One key difference between idiopathic PD and secondary parkinsonism is that idiopathic PD typically responds better to pharmacological dopaminergic treatment compared to secondary parkinsonism (Rizek et al., 2016). However, based on the results in this review, this might not always hold true for tumoral parkinsonism. Most guidelines agree that an MRI or CT should always be a part of PD diagnostics (Arcaya Navarro et al., 1986;Choi et al., 2012;Cicarelli et al., 1999;Ho et al., 2008;Wächter et al., 2011). Furthermore, a tumor patient might also have idiopathic PD. Therefore, methods such as MIBG myocardial scintigraphy or DaTSCAN should be considered for further differentiation (Ogawa et al., 2018). It should, however, be pointed out that dopamine transporter imaging has been pathological in single cases clinically identified as tumoral parkinsonism (Modreanu et al., 2020).
Because the tumor can affect various brain segments, patients with tumoral parkinsonism seldom solely manifest with parkinsonian symptoms. Classic symptoms, such as headaches, seizures, and behavioral changes, typically accompany motor symptoms (Perkins & Liu, 2016).
Furthermore, the parkinsonian clinical picture in tumor cases can vary widely and is often atypical compared to idiopathic PD (Saleh et al., 2021).
Although brain metastases outnumber primary brain tumors by a ratio of 10:1 and occur in as much as 25% of cancer patients (Saha et al., 2013), metastases have rarely been reported to cause tumoral parkinsonism. This discrepancy might exist because brain metastases of cancer with extracranial origin usually develop in the later stages of neoplastic disease when other neurological deficits also manifest, thus confounding symptoms. There was no major difference in symptomatology between brain metastasis and primary brain tumor patients.
The general pattern across all these cases is that there is a lot of uncertainty as to how the lesion causes parkinsonian symptomatology. Several explanatory theories have been proposed, including distant mass effect, vascular dysfunction, and nigrostriatal influence, but most studies have not been able to motivate these speculative pathophysiological causalities.
PNSs are complex abnormal immunological reactions to neoplasms and have been reported to cause secondary parkinsonism (Golbe et al., 1989;Li, 2019;Oliveras et al., 1988;Topcular et al., 2013). Recent research indicates that these immune mechanisms may also affect the pathogenesis in idiopathic PD (Chaná-Cuevas et al., 2020;Xing et al., 2022), as microglia, T-cells, and cytokines have been shown to be abnormally increased in these patients (Rocha et al., 2018;Xing et al., 2022). Although not included in the systematic search, Stiff person syndrome deserves to be mentioned in this immunological context. It is a rare autoimmune movement disorder, clinically characterized by fluctuating rigidity and stiffness of the axial and proximal lower limb muscles, with superimposed painful spasms and continuous motor unit activity on electromyography. In 5% of patients with Stiff person syndrome, the underlying cause is PNS. So forth, PNS can cause several movement disorders including parkinsonism and Stiff person syndrome (Hadavi et al., 2011). This connection between paramalignancy, autoimmunity, antibodies, and movement disorders is still heavily underexplored and demands more attention. Oncological treatment has been reported to cause parkinsonism in both adults and children (Artusi et al., 2015;Pranzatelli et al., 1994;Wenning et al., 1999), and while it has been hypothesized that white matter diffuse damage related to radiotherapy might cause parkinsonism (Franchino et al., 2019), the biological mechanisms at play are still not well understood.
Of the total 25 studies, 44% reported no effect, 48% low to moderate effect, and 8% excellent effect on parkinsonian symptoms after trying some sort of dopaminergic therapy (see Table 3 for details).
This indicates that, on the contrary to popular belief, dopaminergic therapy might indeed have potential to relieve symptoms in cases of tumoral parkinsonism. However, the selected analysis approach is gross and simple, and cannot safely determine that there is indeed clinical benefit with dopaminergic treatment. If dopaminergic treatment was indeed to improve symptoms in brain tumor patients with secondary parkinsonism, this would be beneficial to the quality of life of the patient.
Almost all studies included in this review have a major limitation: they neglect nonmotor symptoms. Nonmotor symptoms affect the quality of life immensely (Prakash et al., 2016) and deserve more recognition. Because the included articles do not report sufficient nonmotor data, discussing symptoms in this review, including the fact that dopaminergic treatment might ease them, almost exclusively refers to motor symptoms when not stated otherwise. Since patients with a brain tumor often experience anxiety and depression, there is significant overlap with nonmotor symptoms potentially related to the secondary parkinsonism. The authors therefore suggest that it is important to recognize nonmotor symptoms, and based on clinical experience, that one could consider dopaminergic therapy in tumoral parkinsonism patients with mixed nonmotor symptoms, for example, for improving depression, anxiety, or daytime sleepiness.

Recommended dopaminergic therapy and dosage
Most studies in which dopaminergic treatment was initiated used levodopa (Table 3). Levodopa has likely been the firsthand therapeutic choice for tumoral parkinsonism because of its long history as the most efficacious drug in the treatment of parkinsonism and its relatively few side effects. Several dosages of levodopa have been tried (Table 3), but the data in these studies are not sufficient to make sound conclusions regarding dosage. Treatment results based on specific anatomical location and histological entity are inconsistent and limited. However, cavernomas and adenocarcinoma PNSs might respond poorly, and parkinsonism induced by oncological treatment and brain resection well, to levodopa (Table 3). For the time being, it can probably therefore be recommended that, just like in atypical parkinsonism (Fanciulli et al., 2019), one may increase the dosage of levodopa slowly to a maximum of 1000 mg per day in tumoral parkinsonism before assessing therapeutic effect. Dopamine agonists and amantadine could also TA B L E 3 The impact of dopaminergic treatment on tumoral parkinsonism symptomatology categorized by study, level of effect, prescribed medication, number of patients, and cause of parkinsonism. Postchemotherapy be considered, either in addition to or as an alternative to levodopa.

Study
In the presence of nonmotor symptoms like depression and anxiety, dopaminergic therapy could be tried in addition to the standard treatment.

Limitations of the study
This study regarding secondary tumoral parkinsonism was conducted as a systematic literature review. Articles included were collected through the databases PubMed and Embase. Subject headings such as Emtree terms and MeSH terms were utilized when feasible. Otherwise, free search terms were used, having the additional advantage of finding not-yet-tagged articles. Using the term "astrocytoma" instead of the broader category "glioma" narrowed the search. A majority of the studies did not compare to controls, lacked randomization and blinding, reported single cases, or had small cohorts. Several articles did not report full information regarding which drugs had been used or in what dosage, thus limiting the capacity to analyze the effect of dopaminergic treatment as visualized in Table 3, as well as the results section reporting on chemotherapy.

CONCLUSION
The results of this study suggest that several brain neoplasms, PNSs, and oncological treatments can cause parkinsonism, but that the diverse pathophysiology remains to be fully understood. Published studies indicate that dopaminergic therapy might have relieving effects on motor symptomatology in patients with tumoral parkinsonism.
Dopaminergic therapy has been shown to have a substantial impact on patient quality of life in several cases and should therefore be considered in patients with suspected or confirmed tumoral parkinsonism.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.

FUNDING INFORMATION
No specific funding was received for this work.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.