Diagnostic accuracy of ¹⁸F–FDG PET/CT and MR imaging in patients with adenoid cystic carcinoma

Adenoid cystic carcinoma (ACC) is a rare type of cancer mainly located within secretory glands, most commonly the minor but also the major salivary glands of the head and neck [1]. Each year, about 1200 people are diagnosed with ACC in the United States. ACC is a (generally) well-differentiated and slowly growing tumour with a relatively indolent but relentless course. The tumour is characterised by a marked tendency for perineural invasion with the risk of incomplete resection and (late) local and distant recurrences [2‐4].

Due to the tendency of this tumour to spread along nerve tracts, the complete radical resection of the primary tumour as the main therapeutic approach is very challenging and results in frequent local recurrences. Adjuvant radiotherapy is commonly applied after surgery and leads to improved local control. Furthermore, in some cases, surgery is not feasible to tumour location, therefore primary radiation (e.g. using high linear energy transfer (LET) beams (fast neutrons or carbon ions) [5‐7] is the only treatment option. Chemotherapy has only limited effects in trials so far [8‐10]. Sub-cellular and genetic characteristics of the ACC can be used for targeted systemic therapy efforts and have in part already been considered in clinical trials [11]. Hence, so far no breakthrough has been achieved in systemic therapy for this entity.

In contrast to 10- to 20-year survival rates (about 40% at 15 years), 5-year survival rates in patients with ACC are high with around 89% [12, 13], reflecting a prolonged course of metastatic disease and late local recurrences. Another specific feature of ACC is the relatively rare occurrence of regional lymph nodes metastases although micro metastases might be detected in a relevant number of patients if neck dissection is performed [14]. Even after successful local therapy distant metastases often occur metachronously in lung and sometimes in the liver and limit long-term prognosis. According to the existing literature, distant metastases in ACC are rare at initial diagnosis and depend on disease duration and primary site, although some authors state that current diagnostic means might not be sensitive enough to detect them [15, 16].

Due to the diffuse growth pattern the detection of local recurrence, especially differentiation between scar formation, radiation effects and vital tumour is limited and routine imaging techniques (computed tomography (CT), magnetic resonance imaging (MRI) and ultrasound (US)) suffer from low specificity [17‐19]. Thus, in case of suspected recurrence it is anticipated that metabolic information provided by ¹⁸F–fluorodeoxyglucose (FDG) positron emission tomography (PET) could improve the distinction between benign and malignant lesions [17, 20, 21]. Published reports on FDG uptake in ACC are rare [22‐25], and many of these are only case series with a small number of patients. The existing literature hints that most squamous cell carcinomas are intensely FDG avid, whereas adenoid cystic carcinoma and mucoepidermoid carcinoma show variable uptake depending on the grade of differentiation [26]. In a recently published study by Kim et al. [27] the pretreatment SUVmax of FDG-PET was as a predictor of distant metastasis in ACC of the head and neck.

In this study we evaluate the diagnostic accuracy of FDG PET/CT and MRI for restaging ACC patients in a relatively large population with particular focus on local and whole-body staging.

This study demonstrates the high sensitivity of whole-body FDG PET/CT in detecting both recurrent/residual tumours and regional metastatic spread in patients with ACC. FDG PET/CT also outperformed head neck MRI for local staging and restaging. Moreover, the whole body imaging approach enables the detection of distant metastases even at an early stage and in the follow up.

Local tumours were present in 75% of the patients in our study population. Of these, 63% had residual tumours, and 37% had tumour recurrence. The detection rate of a residual or recurrent tumour on PET/CT was highly accurate with only one false-negative and one false-positive finding. PET/CT was only false-negative in one patient with tumour recurrence in the head and neck area. PET/CT was false-positive in only one case in the head and neck area, possibly due to a physiological hyperfunctional pharyngeal reaction and perfusion after contralateral resection of the primary tumour. MRI had four indistinct findings in the former local tumour region. The diagnostic performance of PET/CT was comparable to MRI in local restaging with a higher sensitivity (96% v. 89%) and diagnostic accuracy (94% vs. 89%) due to less indistinct findings.

Locoregional lymph node metastases were present in only one patient in our study population with a true-positive finding on PET and false-negative finding on CT or MRI due to size criteria. This low incidence of locoregional lymph node metastases is in line with previous reports that described that an ACC rarely metastasises to regional lymph nodes [28, 29]. The ability of PET to detect metastases in lymph nodes that are inconspicuous in morphologic imaging has been shown in a variety of cancers [17, 30, 31].

Combined whole-body PET/CT detected all distant metastases in our study. There were only two cases with disseminated metastases of the lung in which not all of the CT-detectable lesions were also FDG positive due to the limited spatial resolution. In two patients the lung metastases showed a pathologically increased FDG uptake, whereas the CT was indistinctive (differential diagnosis: granuloma). Distant metastases were present in 25% of the patients in our study population. Surprisingly, in the patient group with the primary examination after local tumour resection prior to radiation therapy, 11% of the patients already showed distant metastases in the lung. In three patients secondary carcinomas were found with PET/CT allowing early therapeutic interventions. Previous reports have described overall rate of distant metastases in 25 to 50% of patients with ACC [29], but the occurrence mainly depends on the follow-up period. In fact, in 35 to 50% of patients with ACC, distant metastases were usually found within a follow-up period of more than 15 to 20 years [32]. These studies, which were published several years ago, used conventional diagnostics such as X-ray with comparably low specificity and sensitivity for staging the lung of ACC patients. Therefore, slow-growing metastases might have been misinterpreted as newly occurred metastases at a late time point, although other diagnostic imaging methods like CT might have shown a metastatic disease at an earlier time point in their clinical course or even at time of initial diagnosis.

According to Spiro et al. [29], large primary tumour size and locoregional treatment failure are the most important predisposing factors for outcome. However, distant metastases may develop despite successful locoregional tumour therapy. As in our study, the lungs are the most commonly reported sites to harbour distant metastases. These patients may remain asymptomatic for a long time during their clinical course. However, the onset of symptoms or the development of other visceral metastases usually results in a short survival period. Thus, whole body imaging seems advisable in restaging and follow-up of ACC patients to better understand the course of disease and find optimal therapy strategies, as was recently highlighted in a report from Tewari et al. [33].

In our study, almost all lesions were FDG-avid, and the majority showed a moderately to high FDG uptake. Previously published literature predominantly case reports or case series on FDG uptake in this advancing but relentlessly growing tumour with a propensity for perineural invasion comprise only very few patients [22‐25]. In larger studies ACC often represent only a minority within a group of mixed head and neck cancers [25]. The existing literature of experience with PET imaging show that adenoid cystic carcinoma and mucoepidermoid carcinoma showed variable FDG-uptake depending on the grade of differentiation [26] in comparison to most squamous cell carcinomas that showed high uptake. Due to the complex development of the disease in patients with ACC an appropiate diagnostic imaging is required. The evaluation of changes in tissues and structures in the tumour environment, especially after extended tumour resection in the head and neck area, surgical procedures with transplants or external beam radiotherapy, is a big challenge since US, conventional CT and MRI provide mainly just morphological information and exactly here PET imaging can improve the diagnostic work-up as a tool that depict metabolism information.

A limitation of this study is the nevertheless limited number of patients due to the rarity of this malignant disease. This study cohort comprised 36 patients which is more than any other study. The small sample size did not permit an analysis of the FDG uptake depending on the histopathological ACC subtypes (cribriform, solid and tubular). Another limitation is the reference standard lacking histopathological confirmation. However, the consensus interpretation (adopted by all readers) based on profound imaging data including prior and follow-up examinations (median 12, range 3–105 months).

Interestingly, FDG PET/CT depicted all lesions visualized by the standard imaging tools head and neck is MRI and if applicable thoracic CT. This underlines the potential of whole-body FDG-PET/CT improving the diagnostic accuracy of restaging ACC. Whereas MRI is commonly needed preoperatively for the assessment of tumour delineation and relevant tumour invasion of surrounding structures it is not a routine whole-body imaging technique. The advantages of FDG-PET/CT are the examination of the whole body at one time and at a high sensitivity. Furthermore, our study demonstrates that due to certain weaknesses the conventional CT and MR imaging of the head and neck could benefit from the additional metabolism information provided by the PET component. Thus, in the time of installation of integrated whole-body PET/MR systems, it would be of interest to determine whether a single whole-body PET/MRI could replace a whole-body PET/CT by combining the advantages of both modalities.

Our study demonstrates the potential of FDG-PET/CT as important diagnostic tool in the restaging of ACC patients. FDG PET/CT has a higher sensitivity and diagnostic accuracy than MRI for the loco-regional tumour staging and provides valuable additional information as a whole-body technique.