Evidence-Based PET for Head and Neck Tumours

Head and neck cancer (HNC) accounts for approximately 5% of all malignant tumours with a continuous growing incidence. Head and neck squamous cell carcinoma (HNSCC) represents the majority of HNC [1, 2]. Nodal involvement is frequent in HNC patients, whereas distant metastases are rather uncommon at the time of initial diagnosis and are found approximately in 10% of patients. There is a clear association with lifestyle and factors as alcoholism, smoking, alimentary factors and viruses for the etiological role, while increasing T and N stages remain the most important adverse prognostic factor [3, 4]. Diagnosis of HNC is usually achieved clinically with endoscopy to obtain direct tissue biopsies. Conventional Imaging (CI), including ultrasound (US), computed tomography (CT) and magnetic resonance imaging (MRI) is important for the evaluation of local extension and to provide information about infiltration, involvement of surrounding structures and regional nodal involvement. There is growing evidence, however, that these modalities have limitations for the diagnostic accuracy of nodal involvement and distant metastases. 18F-FDG PET/CT, allowing the analysis of both metabolic and anatomic features, is a very useful imaging tool in HNC, in particular for disease staging, detection of carcinoma of unknown primary (CUP), treatment monitoring, evaluation of residual or recurrent disease and for prognostic information [5–7].


Introduction
Head and neck cancer (HNC) accounts for approximately 5% of all malignant tumours with a continuous growing incidence. Head and neck squamous cell carcinoma (HNSCC) represents the majority of HNC [1,2]. Nodal involvement is frequent in HNC patients, whereas distant metastases are rather uncommon at the time of initial diagnosis and are found approximately in 10% of patients. There is a clear association with lifestyle and factors as alcoholism, smoking, alimentary factors and viruses for the etiological role, while increasing T and N stages remain the most important adverse prognostic factor [3,4]. Diagnosis of HNC is usually achieved clinically with endoscopy to obtain direct tissue biopsies. Conventional Imaging (CI), including ultrasound (US), computed tomography (CT) and magnetic resonance imaging (MRI) is important for the evaluation of local extension and to provide information about infiltration, involvement of surrounding structures and regional nodal involvement. There is growing evidence, however, that these modalities have limitations for the diagnostic accuracy of nodal involvement and distant metastases. 18 F-FDG PET/CT, allowing the analysis of both metabolic and anatomic features, is a very useful imaging tool in HNC, in particular for disease staging, detection of carcinoma of unknown primary (CUP), treatment monitoring, evaluation of residual or recurrent disease and for prognostic information [5][6][7].

Staging
Accurate staging of disease extension at the time of diagnosis is the most important factor for treatment planning and patients prognosis. Furthermore, providing information in early stage of disease is extremely useful for selecting high-risk patients with impact on specifictreatment selection.

T Staging
18 F-FDG PET/CT has high accuracy in detecting the primary tumour but a moderate diagnostic performance than CI to identify the real tumour extension and infiltration of surrounding tissue and structures. These data are necessary for adequate therapeutic strategy and patient prognosis. False-negative results occur on 18 F-FDG PET/CT when the primary tumour is superficial or small, but also in areas of high physiologic activity such as in pharyngeal lymphoid tissue. False-positive results of 18 [9]. Similarly for evaluation of extracapsular spread (ECS), CT and MRI may be similarly effective, whereas evidence was lacking for 18 F-FDG PET/CT and US [10]. 18 F-FDG PET/CT can provide, instead, more useful clinical information and higher sensitivity and specificity (pooled sensitivities and specificity 90% and 89%, respectively) to delineate the presence and extent of mandibular involvement in patients with oral cavity cancer, especially in cases of contextual dental artefacts [11,12]. For evaluation of precancerous and tumour lesions of larynx, Mannelli et al. expressed the need to integrate different imaging methods, proposing a flow chart that allows to stratify patients and select the most appropriate procedure [13]. Overall, the current practice is not in favour of 18 F-FDG PET/CT as gold standard for T staging in HNC in exception of cases with suspect mandibular involvement in oral cavity cancer. The preliminary data about 18 F-FDG PET/MRI demonstrated high sensitivity and moderate specificity of this technique in the diagnosis of HNC lesions, showing also a better tumour delineation. Further investigations are needed to define the real impact of 18 F-FDG PET/MRI in HNC and whether the technique can improve the detection rate of occult primary HNC [14].

Nodal and Distant Metastases Detection
Lymph nodal involvement is the most important prognostic factor in patients with HNSCC with a significant impact on outcome in terms of disease free survival and overall survival. Lymph nodal (N) metastases occur in approximately 50% of HNC patients at the time of diagnosis with a consequent survival decrease. An accurate N staging is therefore a fundamental step. Similarly, the detection of distant metastases at initial staging influences the prognosis avoiding unnecessary radical treatments. Metastases (M) are frequently found in the lungs, followed by the liver and bone. Several data in the literature confirm an excellent diagnostic accuracy of 18 F-FDG PET/CT in N and M staging. A meta-analysis of Vellayappan et al. assessed the diagnostic accuracy of 18 F-FDG PET/CT for staging nasopharyngeal carcinoma (NPC), showing good accuracy of 18 F-FDG PET/CT for N staging (pooled sensitivity and specificity were 84% and 90%, respectively) and for M staging (pooled sensitivity and specificity were 87% and 98%, respectively), but not for T classification [15]. Similarly, Shen et al. confirmed in their meta-analysis an excellent diagnostic performance of 18 F-FDG PET/CT for detecting lymph node and distant metastases in patients with NPC with a pooled sensitivity and specificity of 89% and 96%, respectively [16]. Considering only the detection accuracy for regional nodal metastases in HNC before treatment, 18 F-FDG PET/CT showed good diagnostic performance [17,18]. Moreover, compared with CI, 18 F-FDG PET/CT may have higher per-necklevel sensitivity [19]. These values are even more significant excluding clinically N0 patients with greater accuracy values for 18 F-FDG PET/ CT. Several data showed moderate sensitivity of 18 F-FDG PET/CT for detection of cervical lymph nodal metastases in clinical N0 HNSCC patients with absence of significant better diagnostic accuracy compared to CI; conversely, 18 F-FDG PET/CT has a higher specificity and negative predictive value for the detection of cervical metastatic lymph nodes compared to the other imaging modalities in clinical N0 HNSCC [20][21][22]. Avoiding elective neck dissection is a fundamental step in the diagnostic-therapeutic flow chart of these patients in order to minimize morbidity and health costs. At present elective neck dissection in patients with clinical N0 should not be based upon cross-sectional imaging. A combination of CI and sentinel node biopsy seems to be the preferred staging strategy to reduce the risk of occult metastases in clinical N0 HNSCC [23].
On the other hand, the excellent diagnostic performance of 18 F-FDG PET/CT for detecting distant metastases is clearly underlined in the literature [24][25][26][27][28][29]. Xu et al. showed a pooled sensitivity and specificity of 85.7% and 98.1%, respectively, for 18 F-FDG PET/CT, resulting in a significantly better M staging than CI [26]. This was mainly due to the superior diagnostic performance of 18 F-FDG PET/CT compared to CI in detecting bone metastases [27]. In this setting, 18 F-FDG PET/CT has higher sensitivity compared to bone scintigraphy [28]. On the other hand, for detection of liver metastases 18 F-FDG PET/CT requires further optimization and integration with CI, especially contrast-enhanced CT and MRI [25]. About lung metastases, a metaanalysis demonstrated that 18 F-FDG PET/CT is a valuable diagnostic tool for diagnosing lung malignancies in patients with HNSCC [29].

Prognostic Value
The prognostic value of 18 F-FDG PET/CT has been widely discussed with controversial results. Relevant limiting factors are the variability and reproducibility of each individual parameter. Overall, maximum standardized uptake value (SUVmax), metabolic tumour volume (MTV) and total lesion glycolysis (TLG) were significant prognostic predictors in patients with HNC [30][31][32][33][34][35][36]. No significant correlation was found between metabolic parameters of 18 F-FDG PET/CT in HNC and human papillomavirus (HPV) status [37]. Furthermore, the semi-quantitative PET/CT parameters were not related to histopathological parameters in HNSCC, as Ki67 and p53 [38].

Post-treatment Evaluation
Relevant applications of 18 F-FDG PET/CT in HNC are delineation of the tumour volume for radiation treatment planning, discrimination of post-treatment changes, evaluation of response to multimodality therapy and detection of recurrence.
About radiation therapy planning, Jeong et al. found that 18 F-FDG-avid HNC apparently require 10-30% more radiation dose than FDG-non-avid tumours, supporting radiotherapy boosts for 18 F-FDG-avid tumours; prospective studies are still required in this field [39].
The role of intra-therapy and post-therapy 18 F-FDG PET/CT in predicting long-term survival outcomes in patients treated for HNC has been widely studied. Sheikhbahaei et al. reported that positive results of intra-therapy or post-therapy 18 F-FDG PET/CT could significantly predict the 2-and 5-year risk of death or disease progression [40]. The same group confirmed the high diagnostic performance of 18 F-FDG PET/CT in detecting local, regional and distant recurrences in curatively treated patients with HNC. The pooled sensitivity and specificity of follow-up 18 F-FDG PET/CT for detection of recurrence were 92% and 87%, respectively [41]. These data support its use in clinical practice as confirmed also by other studies that highlight the high accuracy of 18 F-FDG PET/CT performed after the completion of therapy both in NPC and HNSCC before salvage treatment [42][43][44]. 18 F-FDG PET/ CT is also superior to MRI in distinguishing recurrent NPC from fibrosis or scar tissue after radiotherapy in irradiated fields with distortion of normal architecture [45]. Treatment-to-time scan remains a debated aspect. Several works have indicated that early 18 F-FDG PET/CT was less accurate than more delayed imaging after therapy, particularly Cheung and coauthors supported the use of 18 F-FDG PET/CT more than 12 weeks after radiotherapy with or without chemotherapy for the assessment of residual or recurrent HNC [46]. Recently, Helsen et al. confirmed that 18 F-FDG PET/CT performed within 6 months after chemo-radiotherapy in HNSCC patients is the method of choice for ruling out residual/recurrent nodal disease reducing the need for therapeutic intervention [47]. Finally, sensitivity and specificity of 18 F-FDG PET/CT in identifying local failure following curative radiotherapy or surgery for HNSCC were significantly improved when imaging was performed 3 months after end of treatment [48].

Carcinoma of Unknown Origin and Incidental Findings
Several studies have investigated the accuracy of 18 F-FDG PET/CT to identify carcinoma of unknown origin (CUP) in patients with cervical lymph nodal metastases. Generally, the most common sites of detection include the palatine tonsils and the base of the tongue, with increase of falsenegative results when the primary tumour is small or adjacent to physiological uptake sites. Zhu et al. showed a high sensitivity (97%) and a moderate specificity (68%) for the detection of CUP in patients with cervical nodal metastases [49].
Finally, Treglia et al. calculated the pooled prevalence and risk of malignancy of incidental focal 18 F-FDG uptake in the parotid glands. The pooled prevalence of this finding is about 1% of all 18 F-FDG PET/CT. Although these incidental findings are benign in most of the cases, complementary evaluation is needed to exclude malignant lesions or with possible malignant degeneration [50].