SUV semiquantification in PET/CT study By Men chao yang MD.

SUV( 標準攝取值 ) 之重要性 可定義與比較不同時間點之增加或減少 可定義與比較不同時間點之增加或減少 以 ROI 之 SUV 判定惡性之高或低. 效果類似 以 ROI 之 SUV 判定惡性之高或低. 效果類似 以 Tl-201 or Ga-67 secondary or delayed 以 Tl-201 or Ga-67 secondary or delayed imaging, the more SUV ratio in delayed imaging, the more SUV ratio in delayed time, the more malignancy could be reasonably time, the more malignancy could be reasonably suspected. 目前中國大陸大部份醫院以 50-90mins suspected. 目前中國大陸大部份醫院以 50-90mins postinjection scan 為主. 原因 : Benign lesion peak is 30 postinjection scan 為主. 原因 : Benign lesion peak is 30 mins but malignant lesion peak is delayed to 60 mins. mins but malignant lesion peak is delayed to 60 mins. so, 90 mins imaging increased in L/N ratio, and could DDx benign or malignant, increased in accuracy.each bed collection time is 1.5 to 3 mins. SUV, location and configuration of uptake increased in accuracy of diagnosis. Delayed scan 2-3 hrs after this. But both acquisition data should be the same.( 臺灣以 40-60mins 為主 )‏ so, 90 mins imaging increased in L/N ratio, and could DDx benign or malignant, increased in accuracy.each bed collection time is 1.5 to 3 mins. SUV, location and configuration of uptake increased in accuracy of diagnosis. Delayed scan 2-3 hrs after this. But both acquisition data should be the same.( 臺灣以 40-60mins 為主 )‏

2D & 3D collection or scan: Pb or Tungstan as collimation or septa Pb or Tungstan as collimation or septa during a coincidence event detected by during a coincidence event detected by a direct plane or cross plane is 2D. Due to a direct plane or cross plane is 2D. Due to this reason, any length across two detector rings would be stopped by septa angle of collimation. this reason, any length across two detector rings would be stopped by septa angle of collimation. 3 D get rid of collimation. And has more counts than 2D in the same time period. 3 D get rid of collimation. And has more counts than 2D in the same time period. 3 D>2D data and more ring more sensitivity. So 3 D>2D data and more ring more sensitivity. So it needs more correction method. Also decreased scatter effect by get rid of septa. it needs more correction method. Also decreased scatter effect by get rid of septa.

2D analysis of siemens PET

3D different SUV values automatic showed in one click of mouse:

Imaging reconstruction FBP or OSEM FBP or OSEM Using 2D or 3D acquisitions is still under Using 2D or 3D acquisitions is still under debate. Brain imaing use 3D for small activity concentrations outside the true debate. Brain imaing use 3D for small activity concentrations outside the true FOV, could increased sensitivity. FOV, could increased sensitivity. whole body imaging usually done in 2D mode. True FOV in 3D is larger than 2D! whole body imaging usually done in 2D mode. True FOV in 3D is larger than 2D! but hard to reconstruct. High sensitivity. but hard to reconstruct. High sensitivity.

Definition and formula: Normalized for body wt/surface area and injected dose. If really uniformly distributed over entire body, than SUV=1 Normalized for body wt/surface area and injected dose. If really uniformly distributed over entire body, than SUV=1 Formula: Formula: ROI activity(mCi/mL)*body wt(g)/injected dose(mCi)= ROI activity(mCi/mL)*body wt(g)/injected dose(mCi)= mL/g dependent to Lean body mass or body surface mL/g dependent to Lean body mass or body surface area can be substituted for body weight. More accurate for obese patient. area can be substituted for body weight. More accurate for obese patient.

PET/CT SUV is a quantitative value for calculation

SUV could be corrected by…

Eight factors influences: Patient size: 300 lb patient has twice glucose metabolism mass than 150 lb patient. Is it true or false? Obese-much fat-using lean body mass or body surface Patient size: 300 lb patient has twice glucose metabolism mass than 150 lb patient. Is it true or false? Obese-much fat-using lean body mass or body surface area than only use body weight is more area than only use body weight is more accurate in SUV estimation of obese pts. accurate in SUV estimation of obese pts.

Time of measurement Initially 2 hrs, SUV marked increased than Initially 2 hrs, SUV marked increased than after that, slowly increased in SUV. So, after that, slowly increased in SUV. So, early imaging gets low SUV. Delayed scans provide high SUVs. Curve plateaued is the right time of scanning for SUV. The early imaging gets low SUV. Delayed scans provide high SUVs. Curve plateaued is the right time of scanning for SUV. The therapeutic interventions makes the SUV therapeutic interventions makes the SUV plateua earlier. plateua earlier.

Plasma glucose levels. Plasma glucose competition with FDG uptake, so correlation of SUV could be used in this formula: SUV*glucose con/100 mg/Dl Plasma glucose competition with FDG uptake, so correlation of SUV could be used in this formula: SUV*glucose con/100 mg/Dl This is primarily useful for serial monitoring in the same pt, but interstudy is increased in high variability This is primarily useful for serial monitoring in the same pt, but interstudy is increased in high variability

Partial volume effects. Small lesions may have artefactually low Small lesions may have artefactually low SUVs from partial volume effect. SUVs from partial volume effect. appears in< 2-3 FWHM resolution(5- 10mm) in size. But any lesions smaller than 3 cm can potentially show partial volume effect. Less compact tumors(not appears in< 2-3 FWHM resolution(5- 10mm) in size. But any lesions smaller than 3 cm can potentially show partial volume effect. Less compact tumors(not spheric tumors) are more prominent. spheric tumors) are more prominent.

Background activity Spilling in of background activity. Eg. Lung Spilling in of background activity. Eg. Lung tumor has low SUVs in comparison with liver tumor is due to this effect. tumor has low SUVs in comparison with liver tumor is due to this effect.

Dose extravarsation. Underestimated SUV. If already known this, we had better use tumor to background ratio not SUV. Underestimated SUV. If already known this, we had better use tumor to background ratio not SUV.

Reconstruction parameter. AC can affect SUVs. AC can affect SUVs. Filtered back vs interactive reconstruction. Filtered back vs interactive reconstruction. (FBP vs iterative) much different! Numbers of iterations. More its, more SUV (FBP vs iterative) much different! Numbers of iterations. More its, more SUV hot spots. First 5 iterations more increased in SUVs. The maximum SUV will hot spots. First 5 iterations more increased in SUVs. The maximum SUV will increased steadily with more iterations. increased steadily with more iterations. AC like patient motion more effected SUV. AC like patient motion more effected SUV.

CT based AC. Error due to misregistration or truncation Error due to misregistration or truncation artefact is often in this CT based AC. artefact is often in this CT based AC. previously PET scan Ga-68 more accurate previously PET scan Ga-68 more accurate but later PET/CT no difference. Most seen but later PET/CT no difference. Most seen in osseous structures. Now the error is due to different respiratory phases. Lung in osseous structures. Now the error is due to different respiratory phases. Lung SUVs has 30% of variation due to respira. SUVs has 30% of variation due to respira. Truncation due to FOV difference(CT&PET)‏ Truncation due to FOV difference(CT&PET)‏

Standardized SUV ROI in the same institution is very important!avoid err. Automated decreased variability than man. Automated decreased variability than man. 3D isocontour at a % of max. pixel value. 3D isocontour at a % of max. pixel value. Maximum pixel value for small objects to Maximum pixel value for small objects to avoid partial volume errors. Most accurate! avoid partial volume errors. Most accurate! smoothing the data increased PVE. Uptake smoothing the data increased PVE. Uptake is often overestimated at low levels. is often overestimated at low levels.

Standardized SUV ROI in the same institution is very important!avoid err. Fixed-sized ROI centered on maximum pixel value. In different sized tumor to avoid PVE related to size changed. Most Fixed-sized ROI centered on maximum pixel value. In different sized tumor to avoid PVE related to size changed. Most robust ROI method. But inaccurate in small tumors. robust ROI method. But inaccurate in small tumors. Volumetric vs 2D ROI. Choice Volumetric ROI as possible due to 2D has>25% inter Volumetric vs 2D ROI. Choice Volumetric ROI as possible due to 2D has>25% inter observer variability. observer variability.

Maximum vs mean SUV. Max. preferred in a large region of interest Max. preferred in a large region of interest as as areas of necrosis and outside of the as as areas of necrosis and outside of the lesion. But noise will cause Max SUV increased. And max. SUV is less affected by PVE, so its good is smaller lesion. lesion. But noise will cause Max SUV increased. And max. SUV is less affected by PVE, so its good is smaller lesion. Mean SUV be used in a small region of interest. But it should be placed around the most intense area of the region!!! Mean SUV be used in a small region of interest. But it should be placed around the most intense area of the region!!!

SUV cutoff values. Is important in clinical practice or literature. 2.5 is often heard in lung mass. Is important in clinical practice or literature. 2.5 is often heard in lung mass. Advised to use published SUV cutoff values. High or lower is sigficant but not Advised to use published SUV cutoff values. High or lower is sigficant but not the same value—questionable. the same value—questionable. Interpretation:visual uptake;lesion size; Interpretation:visual uptake;lesion size; pattern of uptake; clinical history. No evidence of SUVs is superior than visual.. pattern of uptake; clinical history. No evidence of SUVs is superior than visual..

Therapy response. SUV is a practice method for this purpose. SUV is a practice method for this purpose. Control the factors that influence SUVs. Control the factors that influence SUVs. Calibation between scanner and dose calibator. Longitudinal stability of scanner. Calibation between scanner and dose calibator. Longitudinal stability of scanner. Vigorous QA/QC monitoring procedures. Vigorous QA/QC monitoring procedures. SUVs are most useful in primary diagnosis SUVs are most useful in primary diagnosis if cutoff values are derived from data. if cutoff values are derived from data.

Reporting and Dual time-point imaging Report SUVs of chosen index lesions. Even if these measurements are not used for interpretation of the current study. Report SUVs of chosen index lesions. Even if these measurements are not used for interpretation of the current study. Benign lesions usually decreased over time. Benign lesions usually decreased over time. or remain stable. Early and delayed imaging improve accuracy. Accuracy in thoracic and head and neck tumors. But not in abdominal imaging. Change configuration indicating benign etiology. Time difference should be over 30 mins or more. or remain stable. Early and delayed imaging improve accuracy. Accuracy in thoracic and head and neck tumors. But not in abdominal imaging. Change configuration indicating benign etiology. Time difference should be over 30 mins or more.

Reference PET and PET/CT: A clinical Guide. PET and PET/CT: A clinical Guide. 2n edi Thieme book company. 2n edi Thieme book company. Eugene C. Lin MD. Seattle Washington. Eugene C. Lin MD. Seattle Washington. and Abass Alaci, MD Pennsylvania Uni. and Abass Alaci, MD Pennsylvania Uni. Page Page

BYE-BYE QUESTION OR QUIZ. QUESTION OR QUIZ. which SUV is reliable for smaller lesion? which SUV is reliable for smaller lesion? 1. max 2. mean. 3.min.4. fixed size SUV? 1. max 2. mean. 3.min.4. fixed size SUV? which is correct in whole body imaging? which is correct in whole body imaging? 1. 2D 2. 3D. 3. no collimation? 1. 2D 2. 3D. 3. no collimation? which is used in a large complex tumor with necrosis? 1. max. 2. mean. 3. min.pixel counts? which is used in a large complex tumor with necrosis? 1. max. 2. mean. 3. min.pixel counts? which is more affected by partial volume effect? which is more affected by partial volume effect? 1. large lesion.2.medium lesion.3. small lesion. 1. large lesion.2.medium lesion.3. small lesion.

患者 禁食 6 小时以上，测定空腹血糖 6.2mmol ／ L ，静脉注射显像剂 18F － FDG6.6mCi ，静卧 50 分钟后行全身 PET ／ CT 断层显像，图像同机融合，全身各部位 显像清晰。 患者 禁食 6 小时以上，测定空腹血糖 6.2mmol ／ L ，静脉注射显像剂 18F － FDG6.6mCi ，静卧 50 分钟后行全身 PET ／ CT 断层显像，图像同机融合，全身各部位 显像清晰。 右肺下叶肺门前方见一结节状放射性摄取增高影，大小约 1.8cmX1.8cmX2.4cm ， SUV 最大值为 15.8 ，平均值为 11 ． 7 ， CT 于上述部位见形状不规则结节影；左肺 斜裂见结节状放射性摄取增高影，大小约 0.9cmX1.1cm ， SUV 最大值为 4.2 ，平均值 为 2.1 ， CT 于上述部位见结节影；右肺上叶尖后段见小条状放射件摄取轻度异常增 高影， SUV 最大值为 3.3 ，平均值为 2.1 ， CT 于上述部位见条状阴影；余双肺纹理清 晰，未见实质性病变及异常放射性摄取增高灶。纵隔（ 2R ， 4R 组）及左锁骨上窝 见多个结节状放射性摄取增高影，大小介于 O.7cmXO.7cm ～ 右肺下叶肺门前方见一结节状放射性摄取增高影，大小约 1.8cmX1.8cmX2.4cm ， SUV 最大值为 15.8 ，平均值为 11 ． 7 ， CT 于上述部位见形状不规则结节影；左肺 斜裂见结节状放射性摄取增高影，大小约 0.9cmX1.1cm ， SUV 最大值为 4.2 ，平均值 为 2.1 ， CT 于上述部位见结节影；右肺上叶尖后段见小条状放射件摄取轻度异常增 高影， SUV 最大值为 3.3 ，平均值为 2.1 ， CT 于上述部位见条状阴影；余双肺纹理清 晰，未见实质性病变及异常放射性摄取增高灶。纵隔（ 2R ， 4R 组）及左锁骨上窝 见多个结节状放射性摄取增高影，大小介于 O.7cmXO.7cm ～ 3.0cmX3.9cm 之间， SUV 最大值介于 3.6 ～ 2O.2 之间，平均值介于 2.3 ～ 15.0 之间， CT 于上述部位见多个淋巴结增大影；心脏及大血管正常显影。 3.0cmX3.9cm 之间， SUV 最大值介于 3.6 ～ 2O.2 之间，平均值介于 2.3 ～ 15.0 之间， CT 于上述部位见多个淋巴结增大影；心脏及大血管正常显影。 脑部显影正常，双侧额叶、顶叶、颞叶、枕叶、双侧基底节、丘脑及双侧小脑放 射性分布对称，未见明显异常放射件摄取增高或减低区。 CT 示脑实质内未见异常密 度影，脑沟、脑裂、脑池未见增宽、扩张，其内密度如常。鼻咽部形态正常，未见 放射性异常摄取。颅面部及颈部组织结构、形态及放射性分布未见异常。 脑部显影正常，双侧额叶、顶叶、颞叶、枕叶、双侧基底节、丘脑及双侧小脑放 射性分布对称，未见明显异常放射件摄取增高或减低区。 CT 示脑实质内未见异常密 度影，脑沟、脑裂、脑池未见增宽、扩张，其内密度如常。鼻咽部形态正常，未见 放射性异常摄取。颅面部及颈部组织结构、形态及放射性分布未见异常。

腹膜后见一类圆形软组织肿块影，大小约 3.5cmX4.9cmX4.6cm ，病灶中央见低密度影坏死区， CT 值为 23.1 ～ 36HU ，病灶上界与钩突分界不清， PET 于上述部位见轻度放射性摄取异常增高；膀胱显影正常。前列腺 大小及放射性分布未见明显异常。 腹膜后见一类圆形软组织肿块影，大小约 3.5cmX4.9cmX4.6cm ，病灶中央见低密度影坏死区， CT 值为 23.1 ～ 36HU ，病灶上界与钩突分界不清， PET 于上述部位见轻度放射性摄取异常增高；膀胱显影正常。前列腺 大小及放射性分布未见明显异常。 颈 2 椎体、左侧第 2 侧助、胸 10 椎体及左侧椎板见多个结节状放射性摄取增高影， SUV 最大值介于 3.4 ～ 17.9 之间，平均值介于 2.1 ～ 14.6 之间， CT 于上述部位见骨质破坏；余全身所示骨骼放射性分布未见明显异 常。 颈 2 椎体、左侧第 2 侧助、胸 10 椎体及左侧椎板见多个结节状放射性摄取增高影， SUV 最大值介于 3.4 ～ 17.9 之间，平均值介于 2.1 ～ 14.6 之间， CT 于上述部位见骨质破坏；余全身所示骨骼放射性分布未见明显异 常。 PET/CT 诊断意见： PET/CT 诊断意见： l 、右肺下叶肺门前方结节高代谢病灶，考虑为周围型肺癌；左肺斜裂结节转移。 l 、右肺下叶肺门前方结节高代谢病灶，考虑为周围型肺癌；左肺斜裂结节转移。 2 、纵隔及左锁骨上多发淋巴结转移。 2 、纵隔及左锁骨上多发淋巴结转移。 3 、腹膜后肿块状占位，伴中央坏死，代谢轻度增高，考虑为良性病变。 3 、腹膜后肿块状占位，伴中央坏死，代谢轻度增高，考虑为良性病变。 4 、颈 2 椎体、左侧第 2 侧肋、胸 10 椎体及左侧椎板多发骨转移。 4 、颈 2 椎体、左侧第 2 侧肋、胸 10 椎体及左侧椎板多发骨转移。 5 、右肺上叶尖后段小条状高代谢病灶，考虑为炎症。 5 、右肺上叶尖后段小条状高代谢病灶，考虑为炎症。 6 、肝内多发囊肿；双肾多发结石；脾脏钙化。 6 、肝内多发囊肿；双肾多发结石；脾脏钙化。 7 、前列腺钙化。 7 、前列腺钙化。 肝脏左叶及右叶内散在多个类圆形低密度影，最大约 2.1cmX1.7cm ， CT 值约为 21HU ， PET 于相应部位可见放射 件减低：余肝脏形态、大小、密度及放射性分布未见明显异常。胆囊正常显影，胆囊壁无增厚．未见阳性结石影 。脾脏内见高密度结节状钙化影。胃及十二指肠正常显影。胰腺形态、大小正常，未见放射性异常浓聚影。双肾 盂内见多个高密度结石影，双侧输尿管正常显影。结肠各段见少许生理性摄取，未见明显占位性病变。 肝脏左叶及右叶内散在多个类圆形低密度影，最大约 2.1cmX1.7cm ， CT 值约为 21HU ， PET 于相应部位可见放射 件减低：余肝脏形态、大小、密度及放射性分布未见明显异常。胆囊正常显影，胆囊壁无增厚．未见阳性结石影 。脾脏内见高密度结节状钙化影。胃及十二指肠正常显影。胰腺形态、大小正常，未见放射性异常浓聚影。双肾 盂内见多个高密度结石影，双侧输尿管正常显影。结肠各段见少许生理性摄取，未见明显占位性病变。

患者 男， 38 岁，因 “ 体检发现纵隔内淋巴结肿大十日余 ” 入院。自患病以来无 发热、畏寒，腹胀、腹痛等不适症状。查体：全身浅表淋巴结未触及肿大。实验室 检查未见明显异常指标。 患者 男， 38 岁，因 “ 体检发现纵隔内淋巴结肿大十日余 ” 入院。自患病以来无 发热、畏寒，腹胀、腹痛等不适症状。查体：全身浅表淋巴结未触及肿大。实验室 检查未见明显异常指标。 PET ／ CT 示双侧锁骨上、纵隔 (2 、 4 、 5 、 6 、 7 、 8 组 ) 、双侧肺门、腹腔及腹膜 后多发大小不等的肿大淋巴结影，部分融合成团，最大约 4.60 cm×2.90cm ，放射 性摄取明显增高， SUV 最大值 ，半均 15 ． 2 。 ( 图 1 、 2) 。肝左叶外段见片状放 射性摄取异常增高影，大小 PET ／ CT 示双侧锁骨上、纵隔 (2 、 4 、 5 、 6 、 7 、 8 组 ) 、双侧肺门、腹腔及腹膜 后多发大小不等的肿大淋巴结影，部分融合成团，最大约 4.60 cm×2.90cm ，放射 性摄取明显增高， SUV 最大值 ，半均 15 ． 2 。 ( 图 1 、 2) 。肝左叶外段见片状放 射性摄取异常增高影，大小 约 5.60 cm×3.90 cm ， SUV 最大值 5.60 ，平均 4.80 。 CT 于上述部位见片 约 5.60 cm×3.90 cm ， SUV 最大值 5.60 ，平均 4.80 。 CT 于上述部位见片 状稍低密度影， C T 值 51.10Hu 。 PET ／ CT 初步诊断：①双侧锁骨上、纵隔、双侧 肺门、腹腔及腹膜后多发肿大淋巴结，代谢异常增高，考虑淋巴瘤；②肝左叶高代 谢灶，考虑淋巴瘤浸润。 状稍低密度影， C T 值 51.10Hu 。 PET ／ CT 初步诊断：①双侧锁骨上、纵隔、双侧 肺门、腹腔及腹膜后多发肿大淋巴结，代谢异常增高，考虑淋巴瘤；②肝左叶高代 谢灶，考虑淋巴瘤浸润。 ( 左锁骨上 ) 淋巴结活检：淋巴结肉芽肿性炎，部分结节内可见少量非干酪样坏死 ，倾向于结节病 ( 图 3) 。 “ 强的松 ” 激素治疗 4 个月后，复查 PET ／ CT 见双肺及纵隔淋 巴结明显缩小。 ( 左锁骨上 ) 淋巴结活检：淋巴结肉芽肿性炎，部分结节内可见少量非干酪样坏死 ，倾向于结节病 ( 图 3) 。 “ 强的松 ” 激素治疗 4 个月后，复查 PET ／ CT 见双肺及纵隔淋 巴结明显缩小。