Imaging-Guided Biopsy of. Lesions: Do Lesion Location and Morphologic Features on CT Affect the Positive Predictive Value for Malignancy?

(1)

mize the rate of false-negative biopsies and the need for repeat procedures.

Although the positive predictive value (PPV) of PET for the diagnosis of pulmonary lesions has been well studied [2], its charac- terization in the diagnosis of extrapulmonary lesions is less complete. The vast majority of studies evaluating PET accuracy examined only a single disease or organ system, significantly limiting the generalizability of results [3–16]. Additionally, many studies lacked histopathologic data for all included patients and relied partly on clinical or imaging follow-up to serve as reference standards [4, 8–12, 14–16]. Very few studies have examined the effect of anatomic location on the PET PPV. Studies that have done so were not able to make statistical comparisons because of the sample size [17, 18].

The purpose of our study was to retrospectively determine the effect of lesion location and morphologic appearance on the PPV of

Imaging-Guided Biopsy of

18 F-FDG–Avid Extrapulmonary Lesions: Do Lesion Location and Morphologic Features on CT Affect the Positive Predictive Value for Malignancy?

My-Linh Nguyen^1,2 Debra A. Gervais² Michael A. Blake² Peter R. Mueller² Dushyant V. Sahani² Peter F. Hahn² Raul N. Uppot²

Nguyen ML, Gervais DA, Blake MA, et al.

1Harvard Medical School, Peabody Society, 260 Longwood Ave, Boston, MA 02115. Address correspond- ence to M. L. Nguyen ([email protected]).

2Division of Abdominal Imaging and Intervention, Massachusetts General Hospital, Boston, MA.

AJR 2013; 201:433–438 0361–803X/13/2012–433

I

maging-guided biopsies allow min- imally invasive tissue sampling for diagnosis. For many years, the impetus to biopsy a lesion was to determine whether lesions that were morphologically suspicious on imaging were indeed malignant. In the past 5 years, ¹⁸F-FDG PET/

CT has made inroads in oncology as an imaging tool to identify lesions that are potentially malignant not only on the basis of abnormal morphologic appearance but also on the basis of increased metabolic function. However, because uptake of FDG is an imprecise marker for malignancy, false-positive PET results can be produced by a variety of benign conditions [1]. In determining the likelihood of malignancy of an FDG-avid lesion, it is often helpful to assess the lesion’s location and morphologic appearance on CT. In patients with multiple FDG-avid lesions, stratifying lesions by likelihood of malignancy may help identify the highest yield targets for biopsy so as to mini-

Keywords: ¹⁸F-fluorodeoxyglucose, imaging-guided biopsy, interventional radiology, nuclear medicine, PET/CT DOI:10.2214/AJR.12.9166

Received April 25, 2012; accepted after revision September 26, 2012.

OBJECTIVE. The purpose of our study was to analyze the effect of lesion location and morphologic appearance on CT on the positive predictive value (PPV) for malignancy of all extrapulmonary lesions that were ¹⁸F-FDG avid on PET/CT and that were biopsied under imaging guidance.

MATERIALS AND METHODS. Between January 2004 and December 2010, 227 pa- tients underwent imaging-guided biopsy of 231 PET-positive extrapulmonary lesions with diagnostic pathologic results. The PET PPV for malignancy was retrospectively calculated and stratified according to lesion location and morphologic appearance.

RESULTS. The overall PET PPV for malignancy was 72%. Inflammatory processes ac- counted for the majority of benign biopsy results. Lesion location significantly affected the PPV (p < 0.001). Bone (96%) and liver (90%) lesions had significantly higher PPVs for malig- nancy compared with other locations, whereas lymph nodes (60%) had a significantly lower PPV for malignancy. Lesions that were morphologically suspicious and morphologically benign according to CT findings alone were associated with PPVs of 74% and 57%, respectively (p = 0.05). FDG-avid subcentimeter lymph nodes (n = 8) had a PPV for malignancy of 38%.

CONCLUSION. Over half of PET-avid morphologically benign–appearing lesions and one third of PET-avid subcentimeter lymph nodes were found to be malignant at biopsy, suggesting that benign morphologic appearance alone should not preclude further workup of a PET-positive lesion. Biopsies of FDG-avid lesions in liver and bone yielded high rates of true malignancy, whereas biopsies of lymph nodes yielded a lower rate of malignancy compared with other lesion locations.

Nguyen et al.

FDG-Avid Extrapulmonary Lesions and Positive Predictive Value for Malignancy

Vascular and Interventional Radiology Original Research

(2)

PET for all FDG-avid extrapulmonary lesions biopsied under imaging guidance.

Materials and Methods Study Population

This study was compliant with HIPAA and received institutional review board approval with waiver of informed consent. Our radiology procedure database was retrospectively searched for all PET-positive lesions biopsied under imaging guidance. Pulmonary lesions, defined as lesions located within the lungs, mediastinum, or hila, were excluded. Lesions were considered PET positive if they qualitatively showed abnormal FDG uptake in the last PET study preceding biopsy. Assess- ment of FDG uptake was performed by one of five nuclear medicine imagers with 9–20 years’ experience, in a consensus clinical readout of cases with an abdominal radiologist. Between January 2004 and December 2010, 248 PET-positive extrapulmonary lesions were biopsied under imaging guidance in 244 patients at our tertiary care institution.

Seventeen of these lesions did not have diagnostic tissue pathologic results and were excluded. A to- tal of 231 lesions in 227 patients (102 men and 125 women; median age, 62 years; range, 15–91 years) were therefore included in our study. Most patients had a history of malignancy (n = 186; 81.9%), with the most common diagnoses being lung can-

cer (n = 41), lymphoma (n = 37), colorectal can- cer (n = 19), breast cancer (n = 19), and melanoma (n = 11). Lesions were most commonly located in lymph nodes (n = 72), the liver (n = 39), bone (n = 26), and the adrenal glands (n = 26) (Table 1).

PET/CT Imaging

All PET/CT was performed on either a 16- or 64-MDCT hybrid PET/CT system (Biograph 16 or 64, Siemens Healthcare) or a GE Healthcare Dis- covery 64-MDCT scanner. A dose of 15 mCi (555 MBq) of FDG was administered IV approximate- ly 45 minutes to 1 hour before scanning. Initially, a low-dose CT was performed for attenuation correction. Subsequently, a diagnostic contrast mate- rial–enhanced CT was performed after PET/CT using the following parameters: section thickness, 5 mm; table feed, 15 mm/s; pitch, 1.5; tube voltage, 120 kVp; and effective tube current–time product, 200 mAs. Images were reconstructed with 2-mm section thickness at 2-mm intervals.

Biopsy Procedure

Biopsies were performed under conscious seda- tion under CT or ultrasound guidance. In most cases, a coaxial needle (Temno, Cardinal Health) was inserted, and four fine-needle aspirates and three or four core samples were obtained. Lesions were biopsied an average of 20.8 days (median, 14.5 days;

range, 1–207 days) after PET. The modality of im- aging guidance (CT, n = 191; ultrasound, n = 40) and type of tissue specimen obtained (fine-needle aspirate only, n = 18; core samples only, n = 49; or both specimen types, n = 164) were determined by the interventional radiologist performing the biopsy. Complications were reported according to So- ciety of Interventional Radiology clinical practice guidelines [19], with 177 patients experiencing no complications and 50 patients experiencing minor complications requiring no therapy (small bleeds or hematomas, n = 49; small pneumothorax, n = 1).

Pathologic Correlation

Lesions were classified as malignant or benign according to tissue pathology reports. The classification took into account results of repeat biopsy procedures, which were performed on 14 lesions where the initial biopsy was nondiagnostic or sus- pected to be falsely negative.

Morphologic Evaluation

To determine whether lesions appeared suspicious for malignancy on the basis of morphologic characteristics alone, radiology dictation reports of CT studies performed before or at the time of biopsy were reviewed for the presence of the following: lymph nodes with a short-axis diameter of 1.0 cm or greater; new or enlarging sclerotic, lytic, TABLE 1: Positive Predictive Value (PPV) of ¹⁸F-FDG PET Imaging According to Lesion Location

Lesion Location No. of Lesions

Lesion Malignancy Status

PET PPV, % (95% CI) p Bonferroni p

Malignant Benign

All locations 231 166 65 71.9 (65.7–77.3) NA NA

Lymph node 72 43 29 59.7 (48.2–70.3) 0.0058 0.035

Liver 39 35 4 89.7 (76.4–95.9) 0.0065 0.039

Bone 26 25 1 96.2 (81.1–99.3) 0.0035 0.021

Adrenal gland 26 16 10 61.5 (42.5–77.6) NS NS

Intraabdominal or intrapelvic mass 25 19 6 76.0 (56.6–88.5) NS NS

Extraabdominal mass 19 14 5 73.7 (51.2–88.2) NS NS

Chest wall 5 4 1 NA NA NA

Peritoneum 4 1 3 NA NA NA

Mesentery 4 3 1 NA NA NA

Abdominal wall 2 0 2 NA NA NA

Kidney 2 1 1 NA NA NA

Perinephric space 2 2 0 NA NA NA

Omentum 1 1 0 NA NA NA

Pancreas 1 1 0 NA NA NA

Spleen 1 1 0 NA NA NA

Thyroid 1 0 1 NA NA NA

Uterus 1 0 1 NA NA NA

Note—PPVs were calculated only for locations with more than five lesions. p values refer to difference in PPV for lesions within a location group compared with all lesions outside of the group. NA = not applicable, NS = not significant.

(3)

mixed, or ground-glass bone lesions; adrenal nodules with unenhanced attenuation greater than 10 HU; or new or enlarging masses not characterized as benign. Lesions identified in the dictation report as clearly falling into one of these categories (n = 166) were characterized as morphologically suspicious. Lesions described by the report as having no definitive CT abnormality to correlate with abnor- mal FDG uptake (n = 10) were classified as mor- phologically benign. For all other lesions (n = 55), CT images were reviewed with blinding to PET studies and tissue pathologic results by a radiologist with 8 years of experience in abdominal imaging to characterize each lesion as morphologically suspicious or morphologically benign using the previously described criteria. Overall, 201 lesions were classified as morphologically suspicious on the basis of CT, whereas 30 were classified as morphologically benign.

Statistical Analysis

The PPV of PET was calculated for each of the following groups: all lesions, lesions grouped by location (excluding locations with five or fewer lesions), lesions grouped by morphologic classification, lymph nodes grouped by size, lymph nodes grouped by morphologic classification, and lesions grouped by patient malignancy history. The PPV was defined as the proportion of malignant lesions in each group. PPV 95% CIs were calculated using

the Wilson method. The chi-square test (or Fisher exact test for smaller sample sizes) was used to com- pare rates of malignancy between groups. For each location group with more than five lesions, the rate of malignancy within the group was also compared with the rate for all lesions outside of the group via a chi-square test with Bonferroni correction. Loca- tion groups with five or fewer lesions were excluded from this calculation because of the low likelihood that even true differences among them could be identified with statistical significance. Statistical calculations were performed using software (SAS version 9.2, SAS Institute). Differences were con- sidered significant for p values less than 0.05.

Results

Of 231 PET-positive lesions biopsied under imaging guidance, 166 lesions were found to be malignant, and 65 lesions were found to be benign. The overall PPV of PET for malignancy was therefore 71.9% (95% CI, 65.7–77.3%).

The PPV varied significantly with lesion location (p < 0.001), as summarized in Table 1. Lesions located within bone (PPV, 96.2%;

Bonferroni p = 0.021) or liver (89.7%; Bonfer- roni p = 0.039) had significantly higher PPVs compared with lesions in other locations. In contrast, lesions in lymph nodes (59.7%; Bon- ferroni p = 0.035) had a significantly lower PPV compared with lesions elsewhere.

Lesion morphologic appearance also had a significant effect on PPV. Lesions that appeared morphologically suspicious on CT had a significantly higher PPV (74.1%; 149 malignant of 201 lesions) compared with lesions that appeared morphologically benign (56.7%; 17 malignant of 30 lesions; p = 0.047). Among the 17 morphologically benign lesions that were found to be malignant at biopsy (Table 2), the most common lesion location was bone, followed by the adrenal glands (Figs. 1 and 2).

Among 72 lymph nodes biopsied, 64 (89%) measured at least 1.0 cm in short-axis diameter and were automatically classified as morphologically suspicious because of their size.

Of the remaining eight subcentimeter nodes, five appeared morphologically benign on CT, whereas three appeared morphologically suspicious because of factors other than size, such as multiplicity of nodes or hyperenhancement.

Table 3 summarizes the morphologic classification and biopsy outcome of each of the eight subcentimeter nodes. Three of the eight subcentimeter nodes were found to be malignant at biopsy, yielding a PPV of 37.5%. This PPV was less than that of lymph nodes measuring at least 1.0 cm in short-axis diameter (62.5%;

40 malignant of 64 nodes), but the difference was not statistically significant (p = 0.16). Of the five lymph nodes classified as morphologi-

TABLE 2: Malignant ¹⁸F-FDG–Avid Lesions With Benign CT Morphologic Appearance

Patient No. Lesion Location CT Morphologic Appearance Pathologic Diagnosis

1 Bone Normal^a Metastatic squamous cell carcinoma (cervical primary)

2 Bone Normal^a Metastatic small cell carcinoma (lung primary)

3 Bone Normal^a Metastatic adenocarcinoma (lung primary)

4 Bone Normal^a Follicular lymphoma

5 Bone Normal^a Diffuse large B cell lymphoma

6 Bone Normal^a Diffuse large B cell lymphoma

7 Bone Normal^b Diffuse large B cell lymphoma

8 Bone Normal^b Metastatic gastrointestinal stromal sarcoma

9 Bone Subtle sclerosis^a Metastatic small cell carcinoma (lung primary)

10 Adrenal gland Normal (Fig. 1) Intravascular large cell lymphoma

11 Adrenal gland Thickened gland (Fig. 2) Metastatic adenocarcinoma (lung primary)

12 Adrenal gland Thickened gland Metastatic adenocarcinoma (lung primary)

13 Lymph node 1.2 × 0.7-cm internal mammary node (Fig. 3) Metastatic adenocarcinoma (lung primary) 14 Lymph node 1.3 × 0.5-cm supraclavicular node (Fig. 4)^c Recurrent papillary thyroid carcinoma 15 Skeletal muscle Soft-tissue nodule in iliopsoas muscle or tendon Tenosynovial giant cell tumor

16 Skeletal muscle Soft-tissue nodule in paraspinal musculature Metastatic squamous cell carcinoma (lung primary)

17 Spleen 0.8-cm hypodensity Metastatic papillary thyroid carcinoma

aLesion showed abnormal signal intensity on MRI.

bMRI not obtained.

cShowed microcalcifications on ultrasound.

(4)

cally benign, two were found to be malignant (Figs. 3 and 4), yielding a PPV of 40.0%. This PPV was not significantly different from that of morphologically suspicious nodes (61.2%;

41 malignant of 67 nodes; p = 0.35).

The PPV for patients with a history of cancer (73.7%; 140/190 lesions) was not significantly different from that for patients without a histo- ry of cancer (63.4%; 26/41 lesions; p = 0.19).

Overall, 65 of the 231 (28.1%) PET-positive lesions biopsied were found to be benign.

Table 4 shows the pathologic breakdown of these false-positive lesions. Inflammatory processes, including fibrosis, necrosis, and reactive lymphoid hyperplasia, were responsible for the majority of false-positive findings.

Discussion

Our study shows the significant effect of lesion location on the PET PPV. Lesions located in liver or bone had significantly higher PPVs (90% and 96%, respectively) compared with lesions elsewhere, whereas lesions located in lymph nodes had a significantly lower PPV (60%). Although previous studies have attempted to stratify PET PPV by anatomic location [17, 18], to our knowledge, this is the first study of sufficient size to per- mit statistically significant comparisons between location groups. In patients with multiple FDG-avid lesions, such comparisons may aid in biopsy planning by helping identify the lesions most likely to be malignant.

Our observed PPV for hepatic lesions is consistent with previously reported values of 90% by Sahani et al. [20], 96% by Böhm et al. [21], 97% by Ramos et al. [22], and 100%

by Servois et al. [23]. Although previous studies have reported conflicting PPVs for bone lesions [16, 24, 25], our results corrobo- rate those of Pezeshk et al. [25], who evalu- ated a patient population closely resembling our own. In that study, a PPV of 90% was observed for 39 FDG-avid bone lesions biopsied under CT guidance. A number of studies have examined the PPV of FDG-avid lymph nodes; however, because each study focused on a specific region of nodes or a specific primary malignancy, reported PPVs have varied widely, ranging from 43% for local lymph nodes in patients with biliary cancer [26] to 100% for paraaortic lymph nodes in patients with intraabdominal malignancies [27].

PET-positive lesions that appeared morphologically benign on CT were significantly less likely to be malignant than lesions that appeared morphologically suspicious (PPVs of 57% and 74%, respectively). Nevertheless, more than half of benign-appearing FDG-avid lesions were found to be malignant at biopsy, suggesting that benign morphologic appearance alone should not preclude the further workup of a PET-positive lesion. Malignant lesions that appeared benign were most often bone lesions with a normal or subtly sclerotic appearance on CT. However, these bone lesions all had abnormal MRI examinations, when performed.

Subcentimeter lymph nodes were less likely to be malignant compared with nodes measur- A

Fig. 1—56-year-old man with intravascular lymphoma and intensely ¹⁸F-FDG–avid but morphologically normal adrenal glands. Biopsy of left adrenal gland revealed intravascular large B cell lymphoma.

A, Contrast-enhanced CT shows morphologically normal adrenal glands (arrows).

B, PET shows intense FDG uptake within both adrenal glands (arrows).

B

A B

Fig. 2—46-year-old man with non–small cell lung cancer and thickened lateral limb of right adrenal gland, which was PET positive. Biopsy revealed metastatic lung adenocarcinoma.

A, Contrast-enhanced CT shows mild thickening of lateral limb of right adrenal gland (arrow).

B, PET shows ¹⁸F-FDG uptake within right adrenal gland (arrow).

(5)

ing at least 1 cm in short-axis diameter (PPVs of 38% and 63%, respectively). Despite this disparity, more than one third of subcentimeter lymph nodes were found to be malignant at biopsy. This finding corroborates prior reports showing the superior sensitivity of PET and PET/CT in detecting malignant subcentimeter lymph nodes compared with CT alone [28, 29].

We found that the PET PPV was not significantly affected by the presence or absence of a positive malignancy history. Prior studies

by Knight et al. [30] and Lee et al. [31] in the pulmonary literature reported similar findings.

Limitations

Limitations of this study include its retro- spective nature and its exclusion of lesions not biopsied under imaging guidance, including lesions followed by serial imaging or biopsied via non–imaging-guided modalities, such as laparoscopy. Lesions were referred for imaging-guided biopsy at the discretion of clinical

providers, introducing potential selection bias.

Clinically, at our institution, FDG uptake on PET/CT is only assessed qualitatively and not quantitatively. This standard of practice may affect the number of cases identified as having FDG uptake. Furthermore, although the in- clusion of only patients with imaging-guided biopsied lesions provides a useful histopathologic reference standard proof, it also intro- duces potential verification bias.

Conclusion

PET-positive lesions in liver and bone are highly predictive of malignancy. Positive lesions in lymph nodes are relatively less predictive. Benign morphologic appearance on CT should not alone preclude the further workup of a PET-positive lesion, because more than half of PET-avid benign-appearing lesions and one third of PET-avid subcentimeter A

A

B

B TABLE 3: Biopsy Results of ¹⁸F-FDG–Avid Subcentimeter Lymph Nodes

Patient No. Node Location Size (cm) CT Morphologic Classification Pathologic Classification Pathologic Diagnosis

1 External iliac 1.5 × 0.6 Benign Benign Normal

2 Axillary 1.4 × 0.8 Benign Benign Normal

3 Parotid 0.8 × 0.5 Benign Benign Warthin tumor

4 Internal mammary 1.2 × 0.7 Benign (Fig. 3) Malignant Lung adenocarcinoma

5 Supraclavicular 1.3 × 0.5 Benign (Fig. 4)^a Malignant Papillary thyroid carcinoma

6 Parotid 1.0 × 0.8 Suspicious because of hyperenhancement Benign Normal

7 Cervical 1.4 × 0.8 Suspicious because of multiplicity of nodes Benign Reactive hyperplasia 8 Cervical 2.0 × 0.8 Suspicious because of multiplicity of nodes Malignant Breast adenocarcinoma

aAppeared suspicious on ultrasound because of presence of microcalcifications.

Fig. 3—53-year-old woman with non–small cell lung cancer and PET-positive but morphologically benign subcentimeter internal mammary lymph node. Biopsy revealed metastatic lung adenocarcinoma.

A, Contrast-enhanced CT shows morphologically benign internal mammary lymph node (arrow) measuring 1.2 × 0.7 cm.

B, PET shows FDG uptake within left internal mammary lymph node (arrow).

Fig. 4—40-year-old woman with papillary thyroid cancer who presented with small left supraclavicular node, which was PET positive. Biopsy revealed papillary thyroid carcinoma.

A, Contrast-enhanced chest CT shows small 1.3 × 0.5-cm left supraclavicular node (arrow).

B, PET shows ¹⁸F-FDG uptake within left supraclavicular space (arrow).

(6)

lymph nodes are malignant at biopsy. These results are most generalizable to patients who have been referred for imaging-guided biopsy and may aid in counseling patients and clinical providers regarding potential biopsy outcomes, as well as in identifying high-yield biopsy targets in patients with multiple FDG- avid lesions.

References

1. Blake MA, Singh A, Setty BN, et al. Pearls and pitfalls in interpretation of abdominal and pelvic PET-CT. RadioGraphics 2006; 26:1335–1353 2. Gould MK, Maclean CC, Kuschner WG, Rydzak

CE, Owens DK. Accuracy of positron emission tomography for diagnosis of pulmonary nodules and mass lesions: a meta-analysis. JAMA 2001;

285:914–924

3. Metser U, Even-Sapir E. The role of ¹⁸F-FDG PET/CT in the evaluation of solid splenic mass- es. Semin Ultrasound CT MR 2006; 27:420–425 4. Shin DS, Shon OJ, Han DS, Choi JH, Chun KA,

Cho IH. The clinical efficacy of ¹⁸F-FDG-PET/

CT in benign and malignant musculoskeletal tu- mors. Ann Nucl Med 2008; 22:603–609 5. Rebollo Aguirre AC, Ramos-Font C, Villegas

Portero R, Cook GJ, Llamas Elvira JM, Tabares AR. ¹⁸F-fluorodeoxiglucose positron emission tomography for the evaluation of neoadjuvant therapy response in esophageal cancer: systemat- ic review of the literature. Ann Surg 2009;

250:247–254

6. Ciocca V, Miller MC, Keane WM, Bibbo M.

Correlation of positron emission tomography with fine needle aspiration biopsies in head and neck malignancy. Acta Cytol 2010; 54:5–11 7. Collins BT, Gardner LJ, Verma AK, Lowe VJ,

Dunphy FR, Boyd JH. Correlation of fine needle aspiration biopsy and fluoride-18 fluorodeoxyglucose positron emission tomography in the assessment of locally recurrent and metastatic head and neck neoplasia. Acta Cytol 1998; 42:

1325–1329

8. Huebner RH, Park KC, Shepherd JE, et al. A meta-analysis of the literature for whole-body FDG PET detection of recurrent colorectal can- cer. J Nucl Med 2000; 41:1177–1189

9. Zijlstra JM, Lindauer-van der Werf G, Hoekstra OS, Hooft L, Riphagen II, Huijgens PC. ¹⁸F-fluoro-deoxyglucose positron emission tomography for post- treatment evaluation of malignant lymphoma: a sys- tematic review. Haematologica 2006; 91:522–529 10. van Westreenen HL, Westerterp M, Bossuyt

PM, et al. Systematic review of the staging per- formance of ¹⁸F-fluorodeoxyglucose positron emission tomography in esophageal cancer. J Clin Oncol 2004; 22:3805–3812

11. Wong RJ, Lin DT, Schoder H, et al. Diagnostic and prognostic value of [¹⁸F]fluorodeoxyglucose positron emission tomography for recurrent head and neck squamous cell carcinoma. J Clin Oncol 2002; 20:4199–4208

12. Yen RF, Hung RL, Pan MH, et al. 18-Fluoro- 2-deoxyglucose positron emission tomography in detecting residual/recurrent nasopharyngeal carcinomas and comparison with magnetic reso- nance imaging. Cancer 2003; 98:283–287 13. Shim SS, Lee KS, Kim BT, et al. Non-small cell

lung cancer: prospective comparison of integrat- ed FDG PET/CT and CT alone for preoperative staging. Radiology 2005; 236:1011–1019 14. Branstetter BF 4th, Blodgett TM, Zimmer LA, et

al. Head and neck malignancy: is PET/CT more accurate than PET or CT alone? Radiology 2005; 235:580–586

15. Gambhir SS, Czernin J, Schwimmer J, Silver- man DH, Coleman RE, Phelps ME. A tabulated summary of the FDG PET literature. J Nucl Med 2001; 42(suppl 5): 1S–93S

16. Taira AV, Herfkens RJ, Gambhir SS, Quon A.

Detection of bone metastases: assessment of in- tegrated FDG PET/CT imaging. Radiology 2007; 243:204–211

17. Tatsumi M, Miller JH, Wahl RL. ¹⁸F-FDG PET/

CT in evaluating non-CNS pediatric malignan- cies. J Nucl Med 2007; 48:1923–1931

18. Pansare V, Bandyopadhyay S, Feng J, et al. Fine needle aspiration outcomes of masses detected by positron emission tomography: correlation with stan- dard uptake value. Acta Cytol 2007; 51:509–516 19. Sacks D, McClenny TE, Cardella JF, Lewis CA. So-

ciety of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14:S199–S202 20. Sahani DV, Kalva SP, Fischman AJ, et al. Detec-

tion of liver metastases from adenocarcinoma of the colon and pancreas: comparison of mangafo- dipir trisodium–enhanced liver MRI and whole- body FDG PET. AJR 2005; 185:239–246 21. Böhm B, Voth M, Geoghegan J, et al. Impact of

positron emission tomography on strategy in liver resection for primary and secondary liver tumors.

J Cancer Res Clin Oncol 2004; 130:266–272 22. Ramos E, Valls C, Martinez L, et al. Preopera-

tive staging of patients with liver metastases of colorectal carcinoma: does PET/CT really add something to multidetector CT? Ann Surg Oncol 2011; 18:2654–2661

23. Servois V, Mariani P, Malhaire C, et al. Preop- erative staging of liver metastases from uveal melanoma by magnetic resonance imaging (MRI) and fluorodeoxyglucose-positron emission tomog- raphy (FDG-PET). Eur J Surg Oncol 2010; 36:

189–194

24. Tian R, Su M, Tian Y, et al. Dual-time point PET/CT with F-18 FDG for the differentiation of malignant and benign bone lesions. Skeletal Ra- diol 2009; 38:451–458

25. Pezeshk P, Sadow CA, Winalski CS, Lang PK, Ready JE, Carrino JA. Usefulness of ¹⁸F-FDG PET-directed skeletal biopsy for metastatic neo- plasm. Acad Radiol 2006; 13:1011–1015 26. Kobayashi S, Nagano H, Hoshino H, et al. Diag-

nostic value of FDG-PET for lymph node metas- tasis and outcome of surgery for biliary cancer. J Surg Oncol 2011; 103:223–229

27. Lee MJ, Yun MJ, Park MS, et al. Paraaortic lymph node metastasis in patients with intra-ab- dominal malignancies: CT vs PET. World J Gas- troenterol 2009; 15:4434–4438

28. Metser U, You J, McSweeney S, Freeman M, Hendler A. Assessment of tumor recurrence in patients with colorectal cancer and elevated car- cinoembryonic antigen level: FDG PET/CT ver- sus contrast-enhanced 64-MDCT of the chest and abdomen. AJR 2010; 194:766–771 29. Yao M, Smith RB, Hoffman HT, Funk GF, Gra-

ham MM, Buatti JM. Merkel cell carcinoma:

two case reports focusing on the role of fluorodeoxyglucose positron emission tomography imag- ing in staging and surveillance. Am J Clin Oncol 2005; 28:205–210

30. Knight SB, Delbeke D, Stewart JR, Sandler MP.

Evaluation of pulmonary lesions with FDG- PET: comparison of findings in patients with and without a history of prior malignancy. Chest 1996; 109:982–988

31. Lee J, Aronchick JM, Alavi A. Accuracy of F-18 fluorodeoxyglucose positron emission tomography for the evaluation of malignancy in patients presenting with new lung abnormalities: a retro- spective review. Chest 2001; 120:1791–1797 TABLE 4: Benign ¹⁸F-FDG PET–

Positive Lesions (n = 65), by Category

Category No. (%) of

Lesions Inflammatory process 37 (57)

Normal tissue 13 (20)

Benign tumor 8 (12)

Schwannoma 4 (6)

Leiomyoma 1 (2)

Neurofibroma 1 (2)

Hibernoma 1 (2)

Warthin tumor 1 (2)

Adrenocortical adenoma 5 (8) Fibroadipose tissue 1 (2) Macrofollicular thyroid nodule 1 (2)