We assessed the sensitivity of positron emission tomography (PET) for detecting synchronous multiple primary cancers, particularly synchronous esophageal cancers in head and neck cancer patients. We retrospectively reviewed 230 head and neck cancer patients. All the patients routinely underwent the following examinations: urinalysis, occult blood, tumor marker detection [squamous cell carcinoma (SCC), cytokeratin fragment (CYFRA), and carcinoembryonic antigen (CEA)], esophagogastroduodenoscopy, colonoscopy (when CEA was high or occult blood was positive), abdominal ultrasonography, plain chest computed tomography (CT), and PET. Bronchoscopy was performed when CT revealed lung shadow of central region. Synchronous multiple primary cancers were detected in 42 (18.2%) patients. The diagnostic sensitivity of PET for synchronous primary cancers was as follows: esophagus, 7.6% (1/13); stomach, 25.0% (2/8); lung, 66.7% (4/6); head and neck, 75.0% (3/4); colon, 0% (0/1); kidney, 0% (0/1); and subcutaneous, 100% (1/1). The sensitivity of PET for detecting synchronous esophageal cancers is low because these are early-stage cancers (almost stage 0–I). Therefore, it is necessary to perform esophagogastroduodenoscopy for detecting synchronous esophageal cancers. PET is an important additional tool for detecting synchronous multiple primary cancers because the diagnostic sensitivity of PET in synchronous head and neck cancer and lung cancer is high. But PET has the limitation of sensitivity for synchronous multiple primary cancers because the diagnostic sensitivity of PET in synchronous esophageal cancer is very low.