**1. Introduction**

Paraneoplastic syndromes represent a wide range of symptoms associated with malignancy. They are usually systemic dysfunctions which are not a direct result of neoplastic tumor invasion or metastasis. They may manifest with symptoms from various systems and organs, causing endocrine, neurological, dermatological, rheumatological and hematological syndromes, usually, in areas and organs that are not directly affected by the neoplasm. The most common malignancies associated with paraneoplastic syndromes include breast, lung, pancreas, kidney as well as gynecological and hematological tumors. The incidence and prevalence of paraneoplastic disorders is estimated to be 0.89/100,000 person-years and 4.4 per 100,000, respectively [1]. In many cases, the etiology of the syndrome is unknown. Most often, they are autoimmune reactions to circulating tumor antigens or an effect of biologically active substances, such as hormones, growth factors or cytokines secreted by tumor cells.

Paraneoplastic neurologic syndromes (PNS) in the large majority respond well to the symptomatic therapy; however, not in all cases is such treatment effective. The populationbased study estimated an incidence of PNS at about 1/100,000 person-years and a prevalence of 4/100,000, with the incidence ratio increasing over time [2]. The clinical presentation of PNS is excessively variable (Table 1); hence, the proper diagnosis requires extensive workup and clinical experience. In 2004, an international panel of experts recommended diagnostic criteria defining a neurological syndrome as a paraneoplastic syndrome based on the coexistence of onconeural antibodies and the presence of clinical symptoms. Further, they categorized those presentation as "classical" and "non-classical" syndromes. "Classical" syndromes (e.g., Lambert–Eaton myaesthenic syndrome, limbic encephalitis, encephalomyelitis, subacute cerebellar degeneration, sensory neuronopathy, dermatomyositis, or opsoclonus-myoclonus) are more likely to be associated with an underlying malignancy [3].

According to the current knowledge, PNS occur in two main mechanisms of immune response. The first is the direct immune response against neuronal receptors or other cell membrane antigens. The most common types of abovementioned antibodies and their clinical significance are summarized in Table 2. Those syndromes usually present clinically by rapidly developing neurological dysfunction, which might not be related to tumor size or growth and frequently respond well to anti-inflammatory treatment. In the second type, the immunoreactivity is triggered by intra-cellular neuronal proteins, which are more often associated with the development of malignancy. In that mechanism, the immune response targets originate from intracellular antigens/proteins revealed by neuronal death. In those conditions, neurological damage is usually more severe and very often irreversible. The most common types of antibodies against intra-cellular neuronal proteins, associated with underlying malignancy, and their clinical symptoms are summarized in Table 3.

However, tumor removal itself is very seldom linked with good prognosis (e.g., teratoma removal in patients with NMDA (Anti-N-methyl-d-aspartate-mediated encephalitis) [4]), prompt identification and treatment are strongly recommended in order to eliminate proteins triggering the immune response (e.g., small cell carcinoma with anti Hu [5]). Within the pursuit of identifying the origin of malignancy, early imaging using radiological procedures (CT or MRI) often fails to establish the diagnosis. Among the widely available methods of advanced diagnostic procedures, a whole-body [18F]FDG PET/CT might be indispensable in the search for underlying pathology.

In the current state of knowledge, the value of [18F]FDG PET/CT as a pivotal imaging in doubtful cases is recommended but not fully elucidated. The rarity and complexity of neurological PNS leads to a paucity of information necessary to formulate strong diagnostic recommendations.


**Table 1.** Paraneoplastic syndromes of the nervous system classified by location. Adapted from Dalmau J, Rosenfeld MR. Paraneoplastic syndromes of the CNS. *Lancet Neurol*. 2008; 7(4): 327–340. [6].


**Table 2.** Major paraneoplastic onconeuronal antibodies reactive with neuronal membrane antigens. Adapted from Galli, J.; Greenlee, J. Paraneoplastic Diseases of the Central Nervous System [7].

**Table 3.** Major paraneoplastic onconeuronal antibodies reactive with intracellular neuronal antigens. Adapted from Galli, J.; Greenlee, J. Paraneoplastic Diseases of the Central Nervous System [7].


The aim of the study was to assess the usefulness of [18F]FDG PET/CT in patients with high risk of PNS and negative or unremarkable results of conventionalradiological imaging.

#### **2. Methods**

Among the patients diagnosed with neurological syndromes between 2016–2020 in the Neurology Department of University Hospital in Krakow, we selected cases with a clinical picture strongly suggesting an underlying neoplastic background (rapid course of PNS and resistance to conventional first-line treatment).

In 15 of them, with positive or unremarkable results of onconeuronal antibodies, due to negative or unambiguous results of conventional screening (including radiological and endoscopy workup), the [18F]FGD PET/CT was applied.

Among these patients, seven were diagnosed with classical neurological PNS (cerebellar degeneration—3 cases, sensory polyneuropathy—1 case, autoimmune encephalitis—3 cases), and 7 with non-classic PNS (myasthenia gravis (MG)—1, myelitis—1, motor neuron disease—1, sensorimotor polyneuropathy—4). Additionally, a patient with Primary Angiitis of Central Nervous System (PACNS) (1) was included in the analysis, since several cases of sporadic presentation of that disease as a paraneoplastic syndrome are available in the literature.

All patients' demographic data, clinical presentation, cerebrospinal fluid (CSF) evaluation, previous imaging results and onconeuronal antibodies screening were recorded from the available patient database. The distribution of PNS types diagnosed among the analyzed group is presented in Table 4.

**Table 4.** The distribution of PNS types diagnosed among the analyzed group and the results of [18F]FDG examination.


#### *2.1. Image Acquisitions and Analysis*

[18F]FDG PET/CT examinations were conducted in accordance with the standard protocol on GE DISCOVERY 690 VCT scanner (Krakow, Poland). All patients fasted at least 6 h before the procedure. Bedside fasting blood glucose of more than 11.0 mmol/L was not accepted. Imaging was performed 60 min after intravenous administration of the radiotracer. An initial low-dose CT without contrast enhancement (Smart mA with range: 50–180 mA) was performed, in order to correct for photon attenuation and to co-localize FDG uptake and anatomical structures. The FDG PET/CT was performed from the midthigh to the top of the head. Patients received 4 MBq [18F]FDG per 1 kg body mass (2 min per bed position in three-dimensional mode). The PET data were reconstructed with the use of the GE (matrix size 128 × 128, Vue Point FX reconstruction method: 16 subsets, 3 iterations reconstruction algorithm). Corrections were applied for attenuation, dead time, scatter and random coincidences. All image analyses were performed on fused PET/CT data sets. Slice thickness of the PET short-axis images was 3.3 mm. The estimated dose of radiation per patient was about 9.5 mSv.

All images were reviewed using first attenuation-corrected images. In all cases, a senior nuclear medicine consultant made an initial assessment of [18F]FDG-PET scans, which was then re-assessed by the consultant radiologist.

The PET/CT was binary classified, so a positive result was defined as the presence of a metabolically active tumor of any region of body, suspicion of metabolically active dissemination of neoplastic disease or increased focal uptake, being highly suspected to correspond with malignancy.

The negative results were considered as scans with no FDG uptake, and with any focal increased uptake with intensity higher than the surrounding tissues, but in localization not highly suspicious as a malignancy.
