FDG-PET and PET/CT - Part II

Positron emission tomography (PET) with 18 F fluoro-deoxyglucose (FDG) is now an established functional imaging modality predominantly used in the work-up of several neoplastic diseases. It also has several neurological and cardiac applications in routine clinical practice. However, the radiopharmaceutical, 18 F-FDG, most commonly used for clinical PET studies today is also taken up by inflammatory and infectious cells and it also has a potential role in inflammation imaging in the future. Since this technique provides a map of glucose metabolism in the body, it is extremely important to understand the bio-distribution of FDG in the human body and factors that alter it. Accordingly, the technique used and several patient factors have a significant impact on the quality of images obtained. Hence, it becomes critical to perform this highly sophisticated exam with adequate patient preparation, following an accepted technique and interpret the images with the knowledge of normal and physiologic bio-distribution of FDG in several body organs and tissues. With this objective, this two-series review article will review the current principles and practice of clinical FDG-PET. The first section of this article deals mostly with basic aspects of FDG-PET and PET/CT including properties of FDG, PET instrumentation and technique and normal variants.


Introduction
The high diagnostic accuracy of FDG-PET and PET/CT has been well documented for several neoplasms, with a consequent rapid increase in the utilization and acceptance of PET imaging in the clinical sett ing. As reported earlier, [1] initial work with PET that started in the 1980s had already established the value of FDG-PET in neurologic and cardiac applications. [2] FDG-PET became a clinical reality only when its importance in the management of several common malignancies was demonstrated and documented by many investigators in the early and late 1990s. The Warburg Eff ect (increased aerobic glycolysis in cancer cells) is the basic principle underlying the value of FDG-PET imaging in neoplasms. [3] FDG-PET is now routinely utilized in the evaluation of several neoplasms like lymphoma, melanoma, lung cancer, breast cancer, colorectal cancer, to name a few. Its value in several other neoplasms is being explored in the United States of America (US) with the establishment of a National Oncologic PET registry (NOPR). The introduction of dual modality PET/CT imaging in the late 1990s and early part of this century has added more strength to this powerful imaging modality. Today, FDG-PET or PET/CT imaging has become an integral component in the evaluation of numerous neoplasms and has been shown to have a tremendous clinical impact on the management of a signifi cant number of cancer patients. In addition to FDG, other PET radiopharmaceuticals are also being developed for cancer imaging.
This article reviews the current clinical applications of FDG-PET and PET/CT in certain malignancies, specifi cally head and neck cancer, lung cancer, lymphoma and melanoma. Thyroid neoplasms are included in the head and neck cancer section and pleural malignancies are discussed in the lung cancer section.

FDG-PET and PET/CT in oncology
Anatomic imaging modalities like CT scan and MRI have superb structural resolution that is continuously improving with technological innovations. However, these imaging modalities rely signifi cantly (although not exclusively) on size and shape of lesions, in addition to indirect fi ndings such as edema, to diff erentiate benign from malignant disease. Therefore, disease manifestations that occur at the molecular and cellular level, before manifesting as structural changes, are frequently not detected with these modalities. FDG-PET has the capability to depict abnormal metabolic activity before any anatomic change occurs. This is particularly important in patients who undergo therapeutic interventions such as radiotherapy and chemotherapy, aft er which structural changes oft en lag behind metabolic changes, for a signifi cant interval of time. In fact, some lesions may never shrink to normal size or disappear completely in spite of adequate treatment. Functional or metabolic imaging with FDG-PET complements structural imaging and helps overcome some of these major shortcomings. As discussed in the previous article in this series, [1] FDG accumulates at a higher rate in cancer cells and remains metabolically trapped in these cells. This phenomenon is the underlying basis for FDG-PET imaging in oncology. FDG-PET has been shown to have a high sensitivity and negative predictive value for the detection of malignancy. Hence in several neoplasms (with notable exceptions), FDG-PET imaging plays a critical role in initial diagnosis, staging, assessing response to therapy, restaging and assessing recurrence. The common neoplasms that have been approved by Centers for Medicare and Medicaid Services (CMS) in the US to be evaluated by FDG-PET include lymphoma, melanoma, head and neck cancer, lung cancer (except small cell lung cancer), breast cancer, esophageal cancer, colorectal cancer, cervical cancer and thyroid cancer (recurrence). However, FDG-PET has also been shown to be useful in several other neoplasms, such as gastric cancer, uterine and ovarian cancers, testicular cancer, soft tissue sarcomas and others. It is also known that a few cancers like prostate cancer and hepatocellular carcinoma do not show increased FDG uptake in a signifi cant portion of patients; and other neoplasms, including pancreatic cancer, are aff ected by the low specifi city of FDG uptake. Renal cancers are diffi cult to assess on FDG-PET imaging due to physiologic renal excretion of FDG. In May 2006, a National Oncologic PET Registry (NOPR) was established in collaboration with the American College of Radiology Imaging Network (ACRIN), the American College of Radiology (ACR) and the Academy of Molecular Imaging (AMI), in response to the proposal by Centers for Medicare and Medicaid Services (CMS) to expand coverage of FDG-PET studies to indications not currently eligible for Medicare reimbursement. The NOPR received inputs from and is endorsed by, the ACR, the American Society of Clinical Oncology (ASCO) and the Society of Nuclear Medicine (SNM) and will provide important data regarding the utility of FDG-PET in several neoplasms not routinely evaluated with FDG-PET imaging.

Head and neck cancer
Malignancies involving the head and neck region have generally been declining in incidence, but about 68,000 new cases are still estimated to occur in the US in 2007. [4] These are usually associated with alcohol and tobacco use; and other factors like poor nutrition and oral hygiene, immunity disorders, (human papilloma virus) HPV infections and occupational exposure. Head and neck cancers include malignancies of the lip and oral cavity; nasopharynx oropharynx and hypopharynx; larynx; ear, paranasal sinuses and salivary glands. Thyroid neoplasms will be discussed separately in the next section in view of their completely diff erent behavior and consequent therapeutic implications, as compared to the rest of the cancers in the head and neck. In addition, lymphoma and other malignancies like melanoma may also involve the head and neck region, but will be discussed in their respective sections.
The majority of head and neck cancers tend to be squamous cell carcinomas (HNSCC) that show intense FDG uptake. [5] FDG-PET imaging plays an important role in diagnosis (especially in the sett ing of metastasis from an unknown primary site), staging, restaging (crucial for assessing residual/ recurrent disease) and monitoring response to therapy. The advent of combined PET/CT imaging has further strengthened the role of this imaging modality in the overall workup of head and neck cancers because of the intricate anatomy in this region.
The diagnosis of head and neck cancer is usually established on the basis of directed biopsies during an endoscopic evaluation (laryngoscopy, nasopharyngoscopy or panendoscopy). Head and neck cancers are staged as per the Tumor, Node and Metastasis (TNM) staging criteria. Along with clinical evaluation with direct visualization, contrast-enhanced CT and MRI are extremely useful for optimal assessment of the 'T' stage as they provide further structural information as regards tumor extension and involvement of adjacent structures. FDG-PET imaging lacks the spatial resolution for providing such exquisite structural details. Moreover, small lesions with low tumor volumes, superfi cial lesions (<4 mm deep) and very low grade tumors (including carcinoma in situ) are not readily visualized on FDG-PET imaging. [6] However, FDG-PET plays an important role in the detection of locoregional metastatic lymphadenopathy and distant metastatic disease.
Only about 10-15% of head and neck cancers have distant metastases at the time of initial presentation and hence therapy is mainly determined by 'T' and 'N' status in a majority of patients. Traditional anatomical evaluation of nodal involvement in the head and neck region is suboptimal, since nodes may be enlarged as a result of infection/ infl ammation (this is not uncommon in the head and neck region) and normal-sized nodes may frequently be involved with metastatic disease, leading to inaccurate upstaging or downstaging of the disease with conventional imaging methods. FDG-PET (and especially PET/ CT) imaging has had a tremendous impact in improving the nodal staging of head and neck cancers. Numerous studies have documented the high sensitivity (87%), specifi city (94-95%) and accuracy of FDG-PET imaging in assessing nodal involvement in head and neck cancers. [6][7][8][9][10] FDG-PET imaging frequently detects metastatic disease in normal-sized lymph nodes [ Figure 1 ]. However, caution is recommended in N0 disease per PET as micrometastases cannot be detected by FDG-PET imaging and hence the management of such patients should not solely be determined by FDG-PET fi ndings [11] and other techniques like sentinel node mapping or surgical neck dissection should be employed for optimal 'N' staging in such patients. Also, sometimes malignant cervical nodes with large extensive central necrosis can be falsely negative on FDG-PET, with only mild FDG uptake at the periphery or no uptake at all [ Figure 2].
In addition to improved nodal staging, the most important added value of FDG-PET imaging is in the detection of unsuspected distant metastases [ Figure 3] that can lead to dramatic changes in patient management. As seen in [ Figure 4], unsuspected mediastinal nodal metastases (subsequently proven by mediastinoscopy and biopsy) detected on FDG-PET imaging in this patient with head and neck cancer altered the management options signifi cantly and prevented futile attempts at curative treatment. Incorporation of FDG-PET imaging in the initial staging changes management in about 10% of patients suspected to have advanced HNSCC. [12] Due to its high sensitivity and the ability to perform whole- , from a PET study of a 66-year-old man recently diagnosed to have squamous cell carcinoma of the oropharynx. Clinical and CT evaluation staged the patient as T2, N0, M0. The PET scan reveals the large primary lesion in the oropharynx to be intensely FDG avid (dotted arrow) with another left oral cavity lesion (dotted arrow) and a left submandibular nodal focus (arrows). The patient had no visible lesion in the left oral cavity at the time of initial examination but subsequently developed an ulcerative lesion in the left lingual sulcus. The submandibular nodal focus localized to an 8-mm node on the CT scan and was not considered enlarged per CT criteria. The patient also had a moderate right hilar focus (seen on the MIP image) and subsequently developed into a non-small cell lung cancer a year later , from a PET study of a 44-year-old man with right tonsillar cancer that is intensely FDG avid. However, an enlarged hypodense nodal mass (white arrows) posterior to the angle of the right mandible is completely 'cold' on PET but clinically was hard and consistent with metastasis (also confi rmed on subsequent neck dissection) The PET scan not only shows extremely locally advanced lip neoplasm (arrows) but also extensive loco-regional nodal metastases along with multiple pulmonary metastases, bone and bone marrow metastases (dotted arrow points to an intense right sacral ala lesion), thus signifi cantly upstaging the disease stage body scanning, FDG-PET is used in the evaluation of patients presenting with malignant cervical lymphadenopathy without evidence of a primary neoplastic lesion on conventional investigations, including CT scan / MRI and panendoscopy with directed biopsies. In such cases, FDG-PET and PET/CT can help identify the primary malignant lesion [ Figure 5] in about 25% of cases. [13,14] At our center, a retrospective chart review revealed that using FDG-PET, we were able to identify the primary tumor in 61% of patients (14 of 23 cases) who presented with this scenario [15] with a very high negative predictive value (89%). Our higher rate of detection of the occult primary lesion may be att ributable to several factors, including rigorous scan protocol (valium pre-treatment oral and intravenous hydration, 90-min delayed imaging), experienced readers, interpretations tuned to increase sensitivity (but not specifi city) in this situation [ Figure 5] and a relatively small number of patients in this study. Since head and neck cancer patients are also prone to have synchronous neoplasms either in the head and neck region or elsewhere in the body, as well as subsequent development of second primaries (most frequently in  , from a PET study of a 70-year-old man presenting with a right neck mass. CT scan of the neck revealed bilateral enlarged neck nodes that showed squamous cell carcinoma on biopsy. However, the primary lesion remained unknown after clinical, CT scan and panedoscopic evaluation. The PET scan shows asymmetric tonsillar uptake with slightly more intense uptake in the right tonsil (arrows), as compared to the left. Subsequent biopsy confi rmed a grade 2 invasive right tonsillar squamous cell carcinoma the lung), FDG-PET imaging plays an important role in diagnosing these unexpected lesions [16] [ Figure 6].
FDG-PET imaging is useful in therapy planning for patients undergoing radiation therapy with a curative or palliative intent or as neoadjuvant therapy. The increasing implementation of intensity-modulated radiation therapy (IMRT) is well complemented by the additional functional/ metabolic information provided by the FDG imaging as it allows the delivery of maximal radiation dose to the most metabolically active areas of the tumor and more complete inclusion of loco-regional disease with sparing of the uninvolved areas. FDG-PET is very helpful in evaluating the cervical nodal chains in patients with advanced-stage head and neck cancer, since the presence of contralateral cervical lymphadenopathy will necessitate the inclusion of both sides of the neck in the radiation therapy fi eld.
In addition to the above, FDG-PET imaging is very useful in assessing response to therapy and in restaging of head and neck cancers. [17,18] Following surgery or radiation therapy, it is extremely diffi cult to assess the neck with conventional imaging modalities like CT scan/MRI due to infl ammatory changes, fi brosis, edema and alteration of normal structures. Determining whether residual neoplasm is present in the postsurgical/ post-radiated tumor bed is a daunting task. Compared to conventional radiological examination, FDG-PET has a bett er diagnostic accuracy in the assessment of residual or recurrent malignant disease in the posttherapeutic neck, including avoidance of unnecessary planned surgery in patients with a negative PET study. [19] Indeed, at our center, we have on several occasions seen intense residual disease aft er therapy on a post-treatment yearold man with laryngeal cancer, detected incidentally on a PET study (arrow in the left image) for a right lung mass (which was not FDG avid and was proven to be benign). A repeat PET study (middle image) 3 months after completing concurrent chemoradiation therapy shows signifi cant but incomplete response with some residual disease (arrow). However, biopsy failed to reveal residual disease and the patient was followed routinely without additional therapy. Five months later, the patient required emergent admission for a tracheostomy for shortness of breath. Vocal cord biopsy revealed malignant disease and a subsequent PET study (right image) shows signifi cant progression of disease (arrow).
PET scan with a negative biopsy, which has then progressed to a larger recurrent lesion a few months later [ Figure 7]. A negative tissue biopsy aft er a strongly positive posttreatment PET scan can be caused by a sampling error and warrants a closer follow-up rather than routine surveillance. A decrease in the intensity of uptake on a follow-up scan confi rms a false-positive post-treatment PET scan, usually due to infl ammatory changes. However, persistence of a focally intense lesion or increase in the intensity of uptake warrants invasive evaluation. The timing of post-treatment PET scan is very crucial, especially aft er radiation therapy. Although there are no specifi c recommendations, generally a 3-month interval aft er completing radiation therapy is felt to be adequate to assess response to therapy. However, if surgery is needed for therapy completion, it is preferable to perform the surgical intervention much sooner than 3 months aft er radiation since more delay makes surgical dissection diffi cult due to post-radiation fi brosis. Hence at our center, we image such patients about 6-8 weeks aft er completion of radiation therapy. The superior assessment of response to therapy with FDG-PET imaging has facilitated a more conservative approach in management. In patients undergoing combined chemo-radiation therapy with a complete response on the post-treatment FDG-PET scan, there is a trend towards a more watchful approach rather than proceeding with a completion neck dissection in spite of large nodes (>3 cm) at initial staging.
There are several limitations of FDG-PET imaging in the evaluation of head and neck cancer. Although it may detect tumors that may be missed by anatomic imaging (especially in the sett ing of an unknown primary), the suboptimal spatial resolution as compared to direct visualization with endoscopy and CT scan / MRI limits the evaluation of small and superfi cial lesions. Also, lowgrade tumors may be missed on PET if there is signifi cant intense physiologic FDG uptake in an adjacent structure. As described in the earlier article in this series, [1] a number of structures in the head and neck region show physiologic FDG uptake, which can sometimes be quite intense. However, physiologic FDG activity is usually symmetric and the added information from CT images in a dedicated PET/CT scan can further help to discern this uptake as benign/ physiologic. Asymmetric FDG uptake (especially in the tonsils) can cause problems with interpretation. In patients being evaluated for cancers other than in the head and neck region, asymmetric tonsillar activity [ Figure 8] is frequently caused by tonsillar infl ammation or lymphoid hyperplasia (unless it is strikingly intense, in which case it suggests an undetected second primary malignancy in the tonsil). In order to avoid physiologic muscle and brown fat activity that could potentially aff ect the interpretation of the study in the head and neck region, it is necessary to place the patient in a comfortable position and room temperature. In addition, benzodiazepines (diazepam is most commonly being used), help prevent brown fat activity. Chewing gum on one side in the mouth can give rise to unilateral intense FDG uptake in the muscles of mastication. Benign uptake , from a PET study of an 80-year-old man with a left lung cancer that shows increased FDG uptake (dotted arrow). Incidentally noted is increased uptake in the right tonsil (arrows), which was found to be due to chronic infl ammation Takalkar, et al.: FDG PET and PET/CT in infl ammatory/ infectious lesions, Warthin's tumor of the parotids (which can be intensely FDG-avid), radiationinduced infl ammation and osteoradionecrosis need to be kept in mind. There may be variable uptake in the vocal cords caused either by talking/ snoring during the FDG uptake interval or due to unilateral vocal cord impairment or iatrogenic causes like Tefl on injection.

Thyroid malignancies
Approximately 33,500 new thyroid cancer cases are estimated in the US in 2007, with about 1,500 deaths. [4] Thyroid neoplasms are predominantly of follicular or parafollicular cell origin. Thyroid neoplasms of follicular cell origin include papillary carcinomas, which represent 80% of all thyroid neoplasms; follicular carcinomas (second most common); as well as the less common but more aggressive Hurthle cell (3-10%) and anaplastic carcinomas (1-2%). The para-follicular or neuroendocrine-derived calcitoninproducing C cells give rise to medullary thyroid carcinomas (MTCs), which represent 5-10% of thyroid neoplasms, a small portion of which may be associated with multiple endocrine neoplasm (MEN) syndromes. In addition, there are rare thyroid neoplasms such as primary thyroid lymphomas and sarcomas.
Generally, thyroid neoplasms are diagnosed aft er workup for a painless thyroid nodule/ mass. Most malignant thyroid nodules show increased FDG uptake. [20] Usually, well-diff erentiated thyroid neoplasms tend to have lower FDG avidity (but excellent radio-iodine avidity) and , from a PET study of a 42-year-old man, status post-surgical resection and neck dissection for a T2, N1, M0 left retromolar trigone mucoepidermoid carcinoma. The PET scan shows an incidental intense calcifi ed right thyroid lobe lesion (arrow), which was proven to be a Hurthle cell carcinoma on resection. Another incidental intense focus is seen in the pelvis (dotted arrow), which was found to be a tubular adenoma in a colonic polyp.
poorly diff erentiated thyroid malignancies like Hurthle cell [ Figure 9] and anaplastic carcinomas tend to show intense FDG uptake (but poor iodine accumulation). In fact, the highest published standard uptake value SUV on a PET scan (SUV max : 125.1, SUV avg : 66.2) in the literature has been in an anaplastic thyroid cancer. [21] Hurthle cell thyroid cancer has generally low avidity for radio-iodine and FDG accumulation is usually very intense, revealing disease not detected by other imaging methods in about 50% of cases. [22] False negative FDG-PET exams can be seen in patients with iodine-avid Hurthle cell cancers. However, there is no defi ned role for FDG-PET imaging in the initial diagnosis and staging of thyroid malignancies. More frequently, there is incidental detection of focally intense lesions in the thyroid gland on scans performed to evaluate other malignant conditions. Such focal intense thyroid 'incidentalomas' warrant further evaluation, since about 25-30% of these turn out to be thyroid carcinomas. [23,24] Sometimes, these 'incidentalomas' may be quite subtle [as in Figure 10], but still warrant evaluation since they can be well-diff erentiated thyroid carcinomas (as was proven in this case).
Except for anaplastic carcinomas, which are very aggressive rapidly progressing tumors, follicular thyroid neoplasms are treated with thyroidectomy followed by radio-iodine therapy with I-131. Follow-up usually consists of physical examination, serum thyroglobulin (Sr. Tg) levels and I-131 or I-123 whole-body scans to detect recurrence. Wholebody iodine imaging may be negative in 10-15% of patients , from a PET study of a 65-year-old man with a right lung nodule. The PET scan shows increased uptake in the right lung nodule (dotted arrow). Also noted incidentally is a subtle right thyroid lobe abnormality (arrows), which was proven to be a well-differentiated thyroid carcinoma on further evaluation with elevated Tg levels and clinical suspicion for disease recurrence, which can be a result of small volume disease or tumor dediff erentiation. [25] In this scenario, FDG-PET imaging is believed to be useful in detecting recurrent disease in about 70-90% of cases, especially if the Sr. Tg levels are more than 10 µg/dl. [24] However, the eff ect of TSH levels on the ability of FDG-PET to detect disease recurrence is still not clearly defi ned. Recent research suggests that elevated TSH levels (more from exogenous stimulation with recombinant TSH than with endogenous TSH stimulation via thyroid hormone withdrawal) may improve lesion detectability on FDG-PET imaging. [26][27][28] Such intensely FDG-avid lesions [ Figure 11] suggest dediff erentiation of the tumor (possibly causing lack of iodine avidity) and may also indicate the need for additional surgical resection or radiation therapy. FDG imaging is an independent predictor of survival in patients with recurrent thyroid cancer and its prognosis is worse in patients with positive FDG scans. [29] MTCs behave diff erently from follicular thyroid carcinomas. Instead of secreting thyroid hormones, they secrete calcitonin as well as carcinoembryonic antigen (CEA). Primary treatment is surgical resection in the form of thyroidectomy. Radio-iodine therapy or radio-iodine scanning for detection of recurrence is of no use in these patients. MTCs are followed by assessing serum calcitonin and CEA levels. In the sett ing of elevated serum calcitonin and/ or CEA levels, FDG-PET imaging is probably the most accurate modality to detect recurrent disease with a sensitivity and specifi city of around 80% (sensitivity ranging from 40-95% in various studies). [30][31][32] A recent paper from Memorial Sloan Kett ering Center in New York [30] reported that in patients with Sr. calcitonin levels <500 pg/ml, FDG-PET has limited sensitivity in detecting recurrence and achieves meaningful sensitivity only when the levels are more than 1000 pg/ml.

Lung malignancies
In the United States, lung cancer remains the leading cause of cancer mortality for both men and women. [4] Broadly, lung cancers are classifi ed into two major groups: small cell and non-small cell. Non-small cell lung carcinoma (NSCLC) is the more common type (75-80%) and includes several histological subtypes like squamous, adenocarcinoma, large cell, carcinoid, etc. Small cell lung carcinoma (SCLC) is less frequent (20-25%) but is frequently metastatic at the time of diagnosis and has diff erent staging, therapeutic and prognostic implications. There is increased FDG activity in most lung cancers, but bronchoalveolar carcinomas (BAC) and lung carcinoids are traditionally considered non-FDGavid. However, BAC and atypical lung carcinoids have been known to show increased FDG uptake on occasion [ Figure 12].
A solitary pulmonary nodule (SPN) is a frequent indication for a PET scan. Although many incidentally detected SPNs tend to be benign, in a high-risk patient (e.g, history of smoking), they are more likely to be malignant. As many as 20-30% of lung neoplasms initially present as SPNs and about 10-30% of all malignant SPNs are metastases from extrathoracic neoplasms [ Figure 13]. In view of the improved spatial resolution of CT scans, many more and smaller SPNs are routinely being detected. It is oft en a challenge to diff erentiate benign from malignant lesions. Although structural characteristics such as confi guration, size, enhancement and calcifi cation provide signifi cant information, the majority of SPNs remain indeterminate on conventional imaging. Tissue diagnosis by biopsy or aspiration (bronchoscopic or CT-guided) or wedge resection although invasive is frequently necessary to confi rm the true nature of the lesion. However, in high-risk patients with coexisting morbidities like severe chronic obstructive pulmonary disease (COPD), FDG-PET imaging plays an  The right lung nodule shows intense FDG uptake (arrows) consistent with a malignant etiology. Also noted is signifi cant FDG uptake in the cecum and rectosigmoid colon (dotted arrows), for which a colonoscopic evaluation was recommended. The patient was subsequently found to have a large ulcerated rectosigmoid colon adenocarcinoma with tubular adenomas and pan-diverticular disease elsewhere in the colon. After resecting the colon cancer, the solitary lung nodule was also resected and found to be metastatic.
important role in the workup of these lesions.
In general, any SPN with intense FDG uptake is considered to be of neoplastic etiology unless proved otherwise [ Figure  14]. Although prior literature suggests an SUV cutoff of 2.5 (lesions with SUV >2.5 are considered malignant), [33][34][35] this is an oversimplifi cation. Visual comparison of FDG uptake intensity with the mediastinal blood pool or liver activity is consistent with neoplastic disease unless proved otherwise. Tissue diagnosis is however necessary to confi rm this and to exclude a false positive study is felt to be an equally reliable indicator of a neoplastic etiology. Another important consideration in the evaluation of SPNs with FDG-PET imaging is their size. Generally, modern PET and PET/CT scanners have a spatial resolution of about 6-8 mm and lesions less than 6 mm may not be adequately assessed; the intensity/ SUV may not be very high in such sub-centimeter lesions. Conversely, if a tiny lesion, supposedly below the spatial resolution for PET imaging, shows increased FDG activity, it is quite suspicious for a malignant etiology (even though the intensity may be mild).
Several studies have shown FDG-PET to be highly accurate in diff erentiating benign from malignant pulmonary lesions, with a sensitivity of greater than 95% and a specifi city of 75-80%. [36] The sensitivity is aff ected by the lesion size (lower sensitivity for smaller lesions), its location (lower sensitivity for lesions located near the thoraco-abdominal interface because of respiratory motion or for those close to a 'hot' myocardium) and sometimes patient's body habitus (lower sensitivity due to high noise in large/ heavy patients). The high negative predictive value of FDG-PET scan makes it a very useful modality, except in sub-centimeter nodules, for reasons described above. Pure BAC and very low grade or very well-diff erentiated adenocarcinomas may demonstrate low FDG activity as well. Therefore, follow-up with CT scans is recommended for suspicious lung lesions to ascertain their stability. The specifi city varies depending on the population being studied and is lower in areas where sarcoidosis, fungal granulomas (like histoplasmosis or coccidioidomycosis) and tuberculosis are more prevalent [ Figure 15]. False positive lesions may also be seen with active infectious / infl ammatory lesions like necrotizing granulomas [ Figure 16].
Certain PET imaging centers practice dual-time point imaging (at 1 and 1.5-2 h) to improve the specifi city for malignant pulmonary lesions. [37][38][39] The hypothesis of dualpoint imaging is based on studies that show increased FDG accumulation in malignant lesions 60 to 90 min aft er injection, whereas uptake in infectious /inflammatory lesions tends to decrease with time. Early studies showed that dual-time point imaging improves the accuracy of FDG-PET, but considerable overlap between malignant and infectious/ infl ammatory lesions is still seen.
An important caveat relates to the incidental detection of a focal intense lesion in the lung on FDG-PET imaging without a corresponding CT correlate [ Figure 17]. Since the newer CT scanners are very sensitive for the detection of small (<5 mm) lung nodules, in the absence of a true CT lesion, this is usually an artifact and may be due to a faulty injection technique. Such artifacts are typically seen if blood is withdrawn into the injection syringe before injecting FDG (to ensure correct intravenous access). Sometimes, misregistration between the PET and CT images due to respiratory motion (especially at the lung bases / liver dome) may also suggest a lesion on FDG-PET imaging, though a review of the CT scan images will immediately resolve the issue.  Anatomical imaging modalities such as CT scan or MRI are superior to FDG-PET and PET/CT (using no intravenous or oral contrast and a low-dose CT scan for att enuation correction and anatomic localization only) for evaluation of the 'T' stage. However, FDG-PET imaging is superior to CT imaging for 'N' and 'M' staging. FDG-PET has proven to be more accurate in assessing nodal involvement and it can accurately identify metastatic disease in normal-sized nodes on CT. It can also bett er diff erentiate lymph node metastases from reactive, but enlarged nodes as seen on CT. The sensitivity and specifi city of FDG-PET imaging in N3 disease (contralateral mediastinal and ipsilateral or contralateral supraclavicular nodal involvement) is greater than 90% and its overall accuracy for assessing mediastinal nodal involvement is around 85%. However, caution is recommended in N0 disease per PET since micrometastases cannot be detected by FDG-PET imaging and invasive staging of intra-thoracic lymph nodes may yield positive nodal disease. [40] The specifi city is again aff ected by infectious/ infl ammatory conditions like sarcoidosis, histoplasmosis and tuberculosis. FDG-PET imaging is also superior to other imaging modalities for detecting distant metastatic disease outside the brain. An unexpected distant metastatic lesion may be seen in up to 12% of patients [41] [ Figure 18].
FDG-PET imaging can also detect functional impairment of other structures due to lung carcinoma. Bulky metastatic mediastinal adenopathy can frequently compress/ involve the recurrent laryngeal and phrenic nerves. Elevation of the ipsilateral hemi-diaphragm or lack of FDG activity in the ipsilateral vocal cord with physiologic or compensatory increased FDG activity in the contralateral vocal cord indicates ipsilateral phrenic nerve or recurrent laryngeal nerve impairment respectively [ Figure 19].
FDG-PET imaging is presently considered the most accurate noninvasive means of staging a biopsy-proven NSCLC (excluding brain metastases). In addition to its high accuracy, FDG-PET imaging can also help make the diagnostic and staging workup of NSCLC more effi cient and cost-  eff ective. FDG-PET imaging can help to guide diagnostic biopsies and improve the yield from tissue sampling. We frequently see lung masses with intense FDG uptake but with a bronchoscopic workup negative for malignancy. These lesions are subsequently proven to be malignant on additional biopsies in a majority of such cases. Even CTguided biopsies of lung masses may be falsely negative if a necrotic area is targeted during biopsy. Figure 20 shows a large right lung mass that had a bronchoscopic workup, as well as two CT-guided biopsies that were all negative for malignancy. The PET scan clearly showed a large area of central necrosis to be responsible for the negative biopsies and a third CT-guided biopsy from the peripheral portion of the lesion that was intensely FDG avid, fi nally confi rmed a diagnosis of NSCLC. Also, if a distant metastatic lesion is identifi ed on an FDG-PET scan for a lung mass, a biopsy from that site alone can prove a stage IV lung neoplasm, sometimes obviating the need for biopsy of the primary lesion (especially when lung biopsy may be deemed risky, as in patients with severe COPD). In general, a biopsy of the site suggesting the highest stage of the disease as identifi ed FDG avidity of the lung lesion also provides prognostic information. Since FDG uptake correlates with tumor growth rate, doubling time and proliferation, intensely FDG-avid lesions tend to be more aggressive with poorer prognosis. SUV is considered an independent prognostic factor and tumors with SUV >10 have a median survival of less than a year. [42] In patients undergoing curative, palliative or neoadjuvant radiation therapy, FDG-PET also helps in radiation therapy planning. Incorporation of the PET data in the process of determining the radiation therapy ports allows more accurate tumor localization, leading to tighter contouring and smaller biological volume calculation and sparing of uninvolved lung parenchyma. [43] This may potentially translate into bett er-tolerated radiation therapy with less pulmonary side eff ects and also bett er control of the disease by avoiding treatment misses to involved areas, as well as allowing the maximum possible radiation dose to the smallest possible target volume. In addition, PET/CT with respiratory gating (also referred to as '4D PET/CT') may allow even bett er targeting of the tumor by taking into consideration respiratory motion of the lesion in the therapy planning protocol. [44] FDG-PET imaging is also playing an increasing role in restaging lung cancer patients to assess response to therapy (especially aft er radiation therapy) and to detect recurrence. It is superior to CT scan for diff erentiating post-therapy changes (aft er surgery, chemotherapy and/ or radiation therapy) from residual or recurrent malignant disease [ Figure 21] and the additional information from the CT images imparts greater specifi city to the interpretation of the PET fi ndings. The timing of assessment aft er radiation therapy is crucial; and frequently, if performed earlier, there is high uptake in the radiation therapy port. This may persist for a considerable amount of time (3-6 months) and may be quite intense in the presence of frank radiation pneumonitis. Although the intensity of FDG activity is expected to subside in 6 months or so, it may remain elevated for even longer [ Figure 22] and generally stable fi ndings are considered negative for residual or recurrent malignant disease. Increasing intensity, focal intense uptake or a new intense lesion in or around the radiation treatment area are suspicious for recurrence. At our center, we are currently evaluating the changes in FDG-PET fi ndings over a prolonged period of time aft er localized radiosurgery using tomotherapy for lung cancer. An 80% decrease or more in SUV predicts a high likelihood of complete response irrespective of the subtype of NSCLC, type of neoadjuvant therapy and fi nal absolute SUV of the lesion. [45] SCLC also shows intense FDG uptake. However, FDG-PET Figure 21 (A, B): PET studies of a 79-year-old lady with a remote history of right lung cancer, with MIP (A) and axial (B) CT scan (top row) and fused PET/CT (bottom row) images at the level of the right upper lobe mass from three consecutive PET scans. Initial study after radiotherapy (left) shows an intensely FDG-avid right upper lung mass, proven to be recurrent neoplasm. A repeat PET study (middle) 4 months after completing concurrent chemoradiation therapy reveals incomplete response with small-to-moderate amount of intense residual disease. Since post-radiation therapy changes can last and remain intense for a considerable interval of time, the patient was observed without additional therapy. A follow-up PET study (right) 3 months later (7 months after completing chemoradiation) reveals disease progression A B imaging is currently not approved by CMS for the workup of SCLC because it is frequently metastatic at the time of diagnosis and the addition of FDG-PET imaging is not felt to have a clinical impact on patient management and outcome. SCLC is generally staged as limited-and extensive-stage disease; and potentially, a small number of limited-stage SCLC patients (with T1-T2, N0, M0 disease) may be curable with aggressive surgical resection followed by chemotherapy. The major consideration in staging SCLC is to determine the need to add radiation with chemotherapy since limited-stage patients are considered for combined chemoradiation therapy with curative intent. FDG-PET has a potential role in SCLC to confi rm limited-stage disease or upstage these patients to extensive-stage disease. However, this is all still under investigation.
FDG-PET imaging is also useful in evaluating pleural malignancies like mesothelioma [ Figure 23], which tend to show intense FDG uptake. In equivocal cases, pleural eff usions may also be suggestive of malignancy if they show even mild increased FDG activity. Focally intense Figure 22 (A, B): PET studies of a 58-year-old lady with left lung cancer. MIP images (A) and axial fused PET/CT images (B) of the left upper lobe mass of six consecutive PET scans are shown. The initial study (I) shows intense FDG uptake in the left upper lung neoplasm without loco-regional nodal or distant metastatic disease. A repeat study (II) 5 months after completing radiation therapy shows interval improvement with signifi cant residual uptake. Subsequent studies (III-VI) 7 months, 11 months, 15 months and 22 months after completing therapy reveal gradual improvement in the intensity of uptake without evidence of residual/ recurrent disease A B abnormalities within the pleural eff usion/abnormality may help guide biopsy. In the evaluation of lung and pleural malignancies with FDG-PET imaging, it is important to bear in mind that talc pleurodesis shows intense FDG uptake [ Figure 24] that is usually sustained for a prolonged period of time and can sometimes wax and wane over time. Obtaining a relevant clinical history in such cases is obviously of paramount importance for avoiding mistakes in interpretation.

Lymphomas
Lymphoma is the fi ft h most common cancer in the US and is one of the few malignancies whose incidence The study shows a pleura-based malignant process in view of the mildly increased FDG activity along the right lung pleura with scattered, more intense pleura-based nodularities (arrows) and metastatic mediastinal adenopathy. Pleural biopsy confi rmed mesothelioma Figure 24: Thirteen millimeter coronal PET images in a 64-year-old man with an abnormal CT scan of the chest. The patient had a history of talc pleurodesis for recurrent right pleural effusion, which is responsible for the intense pleura-based FDG activity around the right lung is rising. [4] Lymphoma is a heterogeneous group of neoplasms involving the immunological system and has several classifi cations. Broadly, it is classifi ed into two major types: Hodgkin's disease (HD) and non-Hodgkin's lymphoma (NHL), both of which have several subtypes. CT scan is considered the predominant modality of choice for lymphoma workup, although it is likely to be replaced by FDG-PET/CT imaging in the near future. Traditionally, gallium (Ga-67) scintigraphy has been used for staging and restaging lymphoma. However, FDG-PET off ers several advantages over Ga-67 imaging in the form of better dosimetry (lower radiation dose), quicker scans (1.5-2 h vs. 2-3 days or more), higher sensitivity (due to superior spatial resolution, lesser physiologic bowel uptake and bett er contrast resolution), bett er quantifi cation potential (and hence bett er prognostic information) and superior assessment of response to therapy. [46][47][48][49][50][51] FDG-PET imaging has an important role in all aspects of lymphoma workup: initial diagnosis, staging, grading, restaging (evaluation for recurrence or residual disease), assessing response to therapy and surveillance (especially for higher-stage disease initially). Lymphoma is perhaps the most frequent oncologic condition for which FDG-PET imaging is performed.
Tissue diagnosis is the norm in lymphoma workup. Although FDG-PET does not image lymphoma cells per se (or any other cancer cells, for that matt er) but rather images the intracellular accumulation of FDG-6-phosphate (intracellular glucose, for all practical purposes), the patt ern of FDG-avid lesions may suggest a diagnosis of lymphoma when the scan is being performed for some other reason [ Figure 25]. In addition, when a diagnosis of lymphoma is suspected, it may suggest another more accessible site for tissue diagnosis [ Figure 26]. It is important to note that almost all lymphoma types show increased FDG uptake, [52,53]  MIP image from a staging PET study shows widespread involvement, the retroperitoneal mass showing intense FDG uptake (arrow). However, the neck nodes also show intense uptake (dotted arrow); and if this was known earlier, it would have presented a more accessible option for biopsy and confi rmation of the diagnosis including low-grade NHL (unlike Ga imaging). Though MALT lymphomas have been traditionally considered not FDG avid, [54] a signifi cant number (about 60%) have now been reported to show increased uptake. [52,53] The American Joint Committ ee on Cancer (AJCC) has based the staging of lymphoma on the Ann Arbor classifi cation that considers the number of nodal regions involved, the extent of involvement and extranodal involvement along with modifi ers according to the presence or absence of 'B' symptoms. FDG-PET imaging is crucial for staging lymphomas [ Figure 27] and is considered either as good as or bett er than CT scanning. It is superior for extranodal disease detection, especially splenic involvement. [55][56][57] It is complementary to bone marrow biopsy for assessing bone marrow involvement. [58,59] Although it can miss low-grade, diff use involvement and may have limited specifi city for diff use involvement versus hematopoietic stimulation from other reasons, it can detect focal bone marrow involvement that may be missed on bone marrow biopsy and can also help provide a better site for bone marrow biopsy. It generally upstages about 20% of patients (although a wide range is reported in the literature) and aff ects management in a signifi cant number of patients. [59][60][61][62] FDG-PET/CT is considered bett er than CT alone or FDG-PET alone for staging lymphoma and has been shown to be cost-eff ective and accurate when the initial scans have guided further conventional imaging. [63] The intensity of FDG activity generally correlates with the grade of lymphoma on histology and provides prognostic information that may impact management decisions. [64] Aggressive lymphomas usually show an SUV ≥13 in the most intense lesion and indolent lymphomas usually show an SUV of ≤6 in the most intense lesion; though there is signifi cant overlap. [65] Not infrequently, there is incongruence between the histologic grade reported and the intensity of FDG activity and usually FDG-PET imaging correctly predicts the nature of the lesion (especially if the FDG-PET shows an intense lesion suggesting high-grade/ aggressive disease and the histology shows low-grade lymphoma). The patient depicted in Figure 26 had a diagnosis of low-grade lymphoma on histology, but the PET study clearly and correctly showed the aggressive nature of the disease based on the intensity as well as the widespread disease, also involving the left orbital cavity [ Figure 28]. Lesions of two diff erent grades are known to coexist in the same patient, as well [ Figure 29].
Ga-67 scintigraphy and CT scan have been the mainstay for assessing response to therapy in lymphoma patients; and the Chesson classifi cation, based mainly on size criteria for scanning, has been in vogue since 1999. [66] Based on their response to therapy, this system classifi es patients into one  Figure 26. The arrow points towards ocular involvement, although the patient did not have any ocular complaints at the time Figure 29: MIP image from a PET study of a 69-year-old man with NHL. Biopsy from the enlarged right axillary nodes (arrow) showed a low-grade lymphoma. The PET study shows moderate uptake in the bilateral neck; supraclavicular, axillary, mediastinal and inguinofemoral lymphadenopathy consistent with the biopsy-proven low-grade lymphoma. However, there is an additional concurrent intense focus in the right pelvis/ iliac region (dotted arrow), indicating a high-grade neoplastic lesion in the same patient of several groups: Complete Response (CR), Unconfi rmed CR (CRu), Partial Response (PR), Stable Disease (SD) and Progressive Disease (PD). With the advent of FDG-PET, these criteria have been shown to be clearly suboptimal [ Figure 30] and were recently revised -with more reliance on FDG-PET fi ndings. [67] The revisions have increased the number of CRs, eliminated the category of CRu and helped discern differences in progression-free survival (PFS) between CR and PR patients.
FDG-PET imaging can correctly assess response to therapy as early as 7 days aft er the fi rst chemotherapy cycle. Recent  Figure 30. Pre-treatment staging PET study (left image) shows widespread involvement. A repeat PET study after two courses of chemotherapy (middle image) shows disease progression. Another PET study after two cycles of changed chemotherapy shows excellent response -to near-complete remission oncology practice management guidelines recommend an FDG-PET scan aft er two cycles of chemotherapy (around 3 weeks aft er the last treatment or just before the next cycle) to assess response and to then proceed with further therapy based on the results of this scan [68,69] [ Figure 31]. It is important to be aware of the known pitfalls and lesions causing false positives while interpreting these studies (like increased uptake in the thymus due to rebound thymic hypertrophy or infl ammatory uptake and increased uptake in the bone marrow due to cytokine stimulation). Assessing response aft er radiation therapy is more complex than monitoring chemotherapy. The timing of the post-treatment FDG-PET scan is of paramount importance for assessing accurate response and there is still no consensus as regards the optimal timing for FDG-PET imaging aft er radiation therapy. [67] Studies and observations from other neoplasms indicate that FDG-PET imaging may be positive due to radiation-induced infl ammation and this may last for a signifi cant amount of time (many months). At our center, we recommend an interval of at least 3 months aft er completing radiation therapy, before advising a repeat scan. In such cases, the patt ern of uptake (within the radiation therapy port, geometric confi guration), the pre-therapy scan and the clinical information help to improve specifi city.
FDG-PET imaging cannot assess microscopic disease and remission (generally meaning less than 1 billion tumor cells) is not equal to tumor eradication or cure. [70] However, a negative FDG-PET scan at the end of therapy does indicate excellent prognosis with sustained progressionfree survival. Conversely, persistent increased FDG uptake during or at the end of therapy indicates a high risk of relapse. [71,72] Hence the current trend is towards assessing response to therapy with FDG-PET imaging early during chemotherapy and then again at the end of therapy.
FDG-PET also has proven value in assessing disease recurrence and residual disease aft er therapy. [73] A mass may never shrink completely but may be completely inactive based on FDG-PET imaging. Also, a mass may remain considerably large, but only a portion of the mass may be active on FDG-PET. In such cases, FDG-PET fi ndings have been proven to be more accurate than CT scan for predicting residual disease. It should be noted that large masses that were intensely FDG avid prior to therapy might retain faint activity for a long period of time, usually a result of post-therapy infl ammatory changes. In addition, FDG-PET also detects disease recurrence earlier; and in patients presenting initially with higher-stage disease (stage III or IV), surveillance with FDG-PET imaging may be costeff ective for detecting recurrences. [74,75] Thus, FDG-PET imaging plays a crucial role in the workup and management of lymphoma. However, it does have certain limitations. [76] Smaller lesions (<7-8 mm) and certain types of lymphoma lesions (very low grade lymphoma, few MALT and mantle cell lymphomas) may be missed. Certain lesions may be masked (renal lesions due to physiologic excretion of FDG by kidneys, GI tract lesions due to intense GI tract uptake in a few patients, patients with intense brown fat uptake, altered biodistribution of FDG due to hyperinsulinemia or rigorous physical activity prior to the FDG-PET imaging). FDG-PET imaging generally has limited sensitivity for detecting metastatic brain lesions. However, usually cerebral lymphoma lesions are 'hot' [ Figure 32]; and in HIV patients with brain lesions, FDG-PET imaging of the brain has become the test of choice to diff erentiate between toxoplasmosis (not 'hot' on FDG-PET) and lymphoma. [77][78][79] Some other conditions may mimic lymphoma, sarcoidosis being a frequent culprit in our Figure 32: Axial brain images with FDG-PET (top two rows) and MRI (bottom two rows) in a patient with CNS lymphoma. Although it is diffi cult to assess malignant lesions in the brain, CNS lymphoma lesions are usually very intense on FDG-PET imaging; and as shown in this example, sometimes more lesions can be seen with FDG-PET brain imaging as compared to brain MRI (arrows) Figure 33: MIP image from a PET study of a 42-year-old man with widespread intensely FDG-avid lesions involving lymph nodes in the neck, supraclavicular regions, axillae, chest (bilateral hilar and mediastinum), abdomen (para-aortic, retroperitoneal and iliac chain) and inguino-femoral regions as well as the spleen along with multiple subdermal/ subcutaneous lesions (arrows), mimicking lymphoma. Biopsy showed sarcoidosis in the nodes, as well as in the subdermal/ subcutaneous lesions part of the US [ Figure 33]. Nodal uptake due to radiotracer administration infi ltration artifact and intense focal GI tract lesions (like benign lymphoid hyperplasia, infl ammatory or adenomatous polyps) may also cause problems with interpretation.

Melanoma
Melanoma is the most serious of skin cancers and is much more frequent in white-skinned patients. In 2007, an estimated 60,000 new cases are expected in the US, with about 8,000 deaths. [4] Diagnosis requires a biopsy and thorough histopathological examination to assess lesion thickness (Breslow's Index) and depth of invasion (Clark's levels), which are important for staging. Early-stage diseases (stages I and II as per AJCC) are potentially curable, with high 5-year survival rates, but the mortality rate is quite high for advanced disease (especially stage IV).
Melanoma is considered one of the most FDG-avid tumors. [80] Imaging with FDG-PET is considered more accurate than standard diagnostic clinical procedures, including clinical examination, radiography, CT scan, MRI, ultrasound and hepatic enzymes. [81,82] However, FDG-PET imaging has a very limited role in early-stage disease. The 'sentinel node' technique is an established modality with high sensitivity, for assessing regional lymph nodes in such cases; and FDG-PET imaging is not an alternative due to its limited sensitivity in detecting micrometastases and low-volume disease (<78 mm 3 ). [83] Moreover, the incidence of regional or distant metastatic disease in early-stage asymptomatic patients aft er adequate treatment of the primary lesion (wide local excision with sentinel node biopsy and nodal basin dissection if necessary) is low enough that routine evaluation with FDG-PET imaging in such patients is not justified. However, in patients with advanced disease, FDG-PET imaging may serve as an important study for overall assessment of the disease (excluding brain and small pulmonary lesions). Although in a patient with confi rmed widespread metastatic melanoma, additional information about extent of disease by FDG-PET imaging may not have signifi cant impact on patient management and outcome; it does help in certain cases. When patients are being considered for surgical resection of isolated distant metastatic lesions, FDG-PET imaging is used to accurately assess disease extent. Additional lesions identified on FDG-PET imaging make surgical resection unwarranted. FDG-PET may have a clinical impact on management in 53% of patients by either upstaging (more frequently) or downstaging patients. [84] Frequently, unexpected nodal or distant metastases and cutaneous/ subcutaneous lesions are detected on FDG-PET imaging. Sometimes, lesions felt to be defi nite visceral metastases on other conventional modalities (like CT or MRI) may be shown to be benign on PET, thus downstaging these patients and making them candidates for potentially curable treatments. However, caution is recommended with pulmonary and intracranial lesions due to reasons described in earlier sections.
FDG-PET has a more established role in assessing response to therapy (especially with the newer treatment options) [ Figure 34] and in detecting recurrent disease [ Figure  35]. [85] Thus, FDG-PET may provide a cost-eff ective tool in managing high-risk patients of malignant melanoma for initial staging and management planning, assessing response, detecting recurrence and guiding additional  workup. The recent mainstream use of combined PET/CT imaging may further improve the sensitivity and specifi city of this imaging modality.
A key issue in scanning melanoma patients with FDG-PET is whether or not to perform a head-to-toes scan (so-called 'true' whole-body imaging). Although no prospective studies have evaluated the value of including the brain and lower extremities in all patients, our current policy is to include the brain and lower extremities only if they are or have been previously, involved (clinically or on other modalities). In the absence of present or past lesions on the lower extremities, we also include them in the imaging fi eld if the patient has proven or suspected disease in the inguino-femoral nodes. Literature review suggests that in the absence of the above factors, the likelihood of fi nding lesions in the lower extremities is very low (less than 1%) to warrant routine inclusion of these areas in imaging protocols. [86]