Comparison of nuclear imaging techniques and volumetric imaging for the prediction of postoperative mortality and liver failure in patients undergoing localized liver-directed treatments: a systematic review

Espersen, Caroline; Borgwardt, Lise; Larsen, Peter Nørgaard; Andersen, Trine Borup; Stenholt, Louise; Petersen, Lars J.

doi:10.1186/s13550-021-00816-4

EJNMMI Research

Table 4 Details of the diagnostic characteristics and predictive values of preoperative nuclear medicine imaging and computed tomography or magnetic resonance imaging for predicting postoperative liver failure (both combined functional/morphological parameters and separate morphological and functional parameters)

From: Comparison of nuclear imaging techniques and volumetric imaging for the prediction of postoperative mortality and liver failure in patients undergoing localized liver-directed treatments: a systematic review

Author	Cutoff value (variable: value and unit)	LF versus no LF (mean or median, p-value)	Key diagnostic characteristics	Univariable regression analysis (impact, p-value)	Multivariable regression analysis (impact, p-value)	Significant differences in key diagnostic characteristics between NMI and CT/MRI
Chapelle et al. [18]	NMI: TLF: NR (%/min)^a	5.0 versus 6.2 (p = 0.020)	NR	NR (p = 0.027)	NR (NS)	NR
	CT: FLRV%: 40 (%)	49 versus 76 (p < 0.001)	Sens 71%, spec 91%, PPV 41%, NPV 97%, AUC 0.77	For FLRV% < 40: OR 26 (p < 0.001)	NR (NS)
	NMI/MRI combined: eFLRF: 2.3 (%/min/m²)	2.2 versus 4.7 (p < 0.001)	Sens 92%, spec 98%, PPV 92%, NPV 99%, AUC 0.89	For eFLRF < 2.3: OR 836 (p < 0.001)	NR (p = 0.001)
Chapelle et al. [26]	NMI: HBS^BSA: NR (%/min)^a	5.5 versus 6.1 (NS)	AUC 0.652	NR (NS)	NR (NR)	NR
	CT: FLRV%: NR (%)	60.6 versus 80.9 (p < 0.001)	AUC 0.800	NR (p < 0.001)	NR (NR)
	NMI/MRI combined: eFLRF: 2.3 (%/min/m²)	3.3 versus 8.4 (p < 0.001)	AUC 0.843	NR (p < 0.001)	OR 0.35 (p = 0.002)
Cho et al. [14]	NMI: SUV_mean: 2.4	2.1 versus 2.3 (NS)	Sens 100%, spec 32%, PPV 7%, NPV 100%	For SUV_mean ≤ 2.4: OR 7.0 (NS)	NR (NR)	NR
	CT: Predicted remnant hepatic volume: 415.8 (cm³)	488.5 versus 652.3 (NS)	Sens 71%, spec 81%, PPV 16%, NPV 98%	For predicted remnant hepatic volume ≤ 415.8: OR 10.6 (p = 0.006)	NR (NS)
	CT: RFRHV: 0.3	0.4 versus 0.5 (p = 0.007)	Sens 71%, spec 88%, PPV 23%, NPV 98%	For RFRHV ≤ 0.3: OR 18.4 (p = 0.001)	NR (NS)
	NMI/CT combined: TLG_r: 625.6	1067 versus 1491 (NS)	Sens 57%, spec 97%, PPV 44%, NPV 98%	For TLG_r ≤ 625.6: OR 36.5 (p < 0.001)	For TLG_r ≤ 625.6: OR 82.9 (p < 0.001)
de Graaf et al. [10]	NMI: FRL-F: 2.69 (%/min/m²)	2.2 versus 4.3 (p = 0.001)	Sens 89%, spec 87%, PPV 57%, NPV 98%, AUC 0.92	NR (NR)	NR (NR)	NR
	CT: %FRL-V: < 30 (normal liver), < 40 (compromised liver) (%)	35.0 versus 49.7 (p = 0.013)	Sens 78%, spec 80%, PPV 44%, NPV 95%	NR (NR)	NR (NR)
	CT: sFRL: < 30 (normal liver), < 40 (compromised liver) (%)	35.2 versus 49.2 (p = 0.018)	Sens 67%, spec 87%, PPV 50%, NPV 93%	NR (NR)	NR (NR)
Dinant et al. [19]	NMI: FRL-uptake: 2.5 (%/min/BSA)	2.3 versus 4.3 (p = 0.003)	Sens 83%, spec 90%, PPV 56%, NPV 97%, AUC 0.90	NR (p = 0.01)	OR 4.0 (p = 0.03)	NR
Dinant et al. [19]	CT: FRL-volume: NR (%)	42 versus 52 (NS)	AUC 0.65	NR (NR)	NR (NR)	NR
Hayashi et al. [20]	NMI: Marginal FR function: depending on ICG-value (see paper for further information) (%)	NR (NR)	NR	For marginal FR function versus safe FR function: OR 11.0 (p = 0.001)	NR (NR)	NR
Hayashi et al. [20]	CT: Marginal FR volume: depending on ICG-value (see paper for further information) (%)	NR (NR)	NR	For marginal FR volume versus safe FR volume: OR 2.3 (NS)	NR (NR)	NR
Hirai et al. [24]	NMI: [^99mTc]Tc-GSA uptake in the FLR: 25 (%)	Before PVE: 10.0 versus 25.8 (p = 0.02) After PVE: 18.0 versus 38.4 (p = 0.01)	Sens 50%, spec 94%, AUC 0.97^b	NR (NR)	NR (NR)	No
Hirai et al. [24]	CT: Ratio of FLR volume to standard liver volume: 35 (%)	Before PVE: 33.0 versus 38.5 (NS) After PVE: 33.7 versus 46.2 (p = 0.003)	Sens 57%, spec 91%, AUC 0.93^b	NR (NR)	NR (NR)	No
Kokudo et al. [17]^c	NMI: LHL15: NR	0.89 versus 0.93 (p = 0.025)	NR	NR (NR)	NR (NS)	NR
	NMI: HH15: NR	0.58 versus 0.52 (NS)	NR	NR (NR)	NR (NR)
	NMI: R₀: NR (μmole)	0.14 versus 0.18 (p = 0.038)	NR	NR (NR)	NR (NS)
	NMI: [R]₀: NR (μM)	0.63 versus 0.70 (NS)	NR	NR (NR)	NR (NR)
	CT: RPF: NR (%)	32.5 versus 27.1 (NS)	NR	NR (NR)	NR (NR)
	Combined NMI/CT: R₀-remnant: 0.16 (μmole)	0.015 versus 0.024 (p = 0.011)	NR	NR (NR)	Per 0.01 μmole increment: HR 0.82 (p = 0.022)
Nanashima et al. ≠ [32]	NMI: LHL15: 0.85	91.1 versus 93.1 (p = 0.014)	NR	NR (NR)	For LHL15 < 0.85 versus ≥ 0.85: OR 1.4 (NS)	NR
Nanashima et al. ≠ [32]	CT: Volume of resected liver: 50 (%)	45 versus 27 (p < 0.01)	NR	NR (NR)	For volume of resected liver ≥ 50 versus < 50%: OR 7.0 (p = 0.027)	NR
Okabe et al. [30]	NMI: LHL15: 0.93^d	0.92 versus 0.93 (p = 0.0027)	Sens: 88%, spec 96%	NR (NR)	For LHL15 ≤ 0.93: OR 7.4 (p = 0.0082)	NR
	NMI: HH15: NR^d	0.66 versus 0.58 (p = 0.0041)	NR	NR (NR)	NR (NS)
	CT: %FLR: NR (%)^d	50.3 versus 50.6 (NS)	NR	NR (NR)	NR (NR)
Olthof et al. [22]	NMI: Total liver function: NR (%/min)	14.6 versus 16.2 (p = 0.41)	NR	NR (NR)	NR (NR)	NR
	NMI: FLR function: NR (%)	44.7 versus 63.4 (p < 0.01)	AUC: 0.68	NR (NR)	NR (NR)
	NMI: FLR function: 8.5 (%/min) (post hoc cutoff)	5.6 versus 8.7 (p < 0.01)	PPV 36%, NPV 91% ^e, AUC 0.69	For FLR function < 8.5: OR 5.4 (p < 0.01)	For FLR function < 8.5: OR 4.1 (p < 0.01)
	NMI: sFLR function: 2.7 (%/min/m²) (predefined cutoff)	3.1 versus 4.7 (p < 0.01)	PPV 38%, NPV 82% ^e, AUC 0.68	NR (NR)	NR (NR)
	CT: FLRV%: 30 (%) (predetermined cutoff), 28.7% (post hoc cutoff)	43 versus 54 (NS)	Predetermined cutoff: PPV 56%, NPV 83%, AUC 0.60 Post hoc cutoff: PPV 71%, NPV 83%	For FLRV% < 30: OR 5.2 (p < 0.01)	For FLRV% < 30: OR 3.4 (NS)
	CT: sFLRV%: NR (%/m²)	NR (NR)	AUC 0.58	NR (NR)	NR (NR)
	CT: FLRV: NR (mL)	760 versus 986 (NS)	NR	NR (NR)	NR (NR)
	CT: Total liver volume: NR (mL)	2016 versus 1841 (NS)	NR	NR (NR)	NR (NR)
Rassam et al. ≠ [23]	NMI: FRLF: 2.7 (%/min/m²) (predefined cutoff)	NR (NS)	NR	NR (NR)	NR (NR)	NR
	NMI: MUR: NR (%/min)	NR (NS)	NR	NR (NR)	NR (NR)
	CT: FRLV%: NR (%)	NR (NS)	NR	NR (NR)	NR (NR)
Serenari et al. [28]	NMI: FLR-C: 34.5 (%)	30 versus 41 (p = 0.011)	Sens 100%, spec 82%, PPV 50%, NPV 100%	NR (NR)	NR (NR)	NR
	NMI: FLR-F: 1.69 (%/min/m²)	0.94 versus 2.07 (p = 0.011)	Sens 100%, spec 75%, PPV 50%, NPV 100%	NR (NR)	NR (NR)
	NMI: HIBA-i: 14.94 (%)	12.86 versus 23.29 (p = 0.001)	Sens 100%, spec 94%, PPV 80%, NPV 100%	NR (NR)	NR (NR)
	CT: FLR/sTLV: NR (%)	35 versus 42 (NS)	NR	NR (NR)	NR (NR)
	CT: FLR/mTLV: NR (%)	34 versus 41 (NS)	NR	NR (NR)	NR (NR)
	CT: FLR/BW: NR (%)	0.74 versus 0.88 (NS)	NR	NR (NR)	NR (NR)
Serenari et al. [29]	NMI: FLR-F: NR (%/min/m²)	1.72 versus 4.02 (NR)	NR	NR (NR)	NR (NR)	NR
Yoshida et al. [31]	NMI: rLUV_(BSA): 27.0	23.0 versus 33.6 (p < 0.001)^f	Sens 91%, spec 81%, PPV 31%, NPV 99%, AUC 0.89	NR (NR)	NR (p < 0.001)	NR
	NMI: rLUV_(BW): 0.66	NR (NR)	Sens 80%, spec 84%, PPV 24%, NPV 99%, AUC 0.85	NR (NR)	NR (NR)
	NMI: rLUR: 50.0 (%)	NR (NR)	Sens 93%, spec 66%, PPV 15%, NPV 99%, AUC 0.87	NR (NR)	NR (NR)
	NMI: HH15: NR	0.64 versus 0.60 (p < 0.05)^f	NR	NR (NR)	NR (NS)
	NMI: LHL15: NR	0.90 versus 0.91 (NS)^f	NR	NR (NR)	NR (NR)
	NMI: %remnant LF: NR (%)	60.9 versus 75.3 (p < 0.001)^f	NR	NR (NR)	NR (NS)
	CT: % remnant LV: NR (%)	61.5 versus 73.7 (p < 0.05)^f	NR	NR (NR)	NR (NS)
	Combined NMI/CT: rLUV_(LV): 0.21	NR (NR)	Sens 92%, spec 46%, PPV 10%, NPV 99%, AUC 0.73	NR (NR)	NR (NR)
Yumoto et al. [16]	NMI: R0-remnant: 100 (nmol/liver)	62.1 versus 122.2 (p < 0.001)	AUC: 0.97	NR (NR)	NR (NR)	NR
	NMI: [R]0: NR (nmol/l)	412 versus 551 (p = 0.045)	AUC: 0.80	NR (NR)	NR (NR)
	NMI: R0: NR (nmol/liver)	149.8 versus 211.2 (p = 0.047)	NR	NR (NR)	NR (NR)
	NMI: LHL15: NR	0.79 versus 0.87 (p = 0.035)	AUC: 0.74	NR (NR)	NR (NR)
	CT: TLV: NR (mL)	1684.3 versus 1429.7 (NS)	NR	NR (NR)	NR (NR)
	CT: Remnant liver volume: NR (mL)	741.7 versus 854.0 (NS)	NR	NR (NR)	NR (NR)
	CT: RPF: NR (%)	36.9 versus 31.7 (p = 0.046)	NR	NR (NR)	NR (NR)

LF: liver failure; NMI: nuclear medicine imaging; CT: computed tomography; MRI: magnetic resonance imaging; NR: not reported; NS: not significant; OR: odds ratio; AUC: area under the curve; HR: hazard ratio; Sens: sensitivity; Spec: specificity; PPV: positive predictive value; NPV: negative predictive value; TLF: total liver function; FLRV%: the future liver remnant volume as a percentage of total liver volume; eFLRF: estimated future liver remnant function; HBS^BSA: global liver function ([^99mTc]Tc-mebrofenin hepatobiliary scintigraphy clearance divided by body surface area); SUV_mean: mean standardized uptake value; RFRHV: ratio of remnant hepatic volume to preoperative hepatic volume; TLG_r: total glycolysis of the remnant liver; FRL-F: future remnant liver uptake function; %FRL-V: future remnant liver volume as a percentage of total liver volume; sFRL: standardized future remnant liver; FRL: future remnant liver; FR: future remnant; FLR: future liver remnant; PVE: portal vein embolization; LHL15: [^99mTc]Tc-GSA receptor index; HH15: [^99mTc]Tc-GSA clearance index; R₀: total hepatic asialoglycoprotein receptor amount; [R]₀: hepatic asialoglycoprotein receptor concentration; RPF: resected parenchymal fraction; R₀-remnant: total hepatic asialoglycoprotein receptor amount in the future remnant liver; %FRL: future remnant liver volume as a percentage of total liver volume; sFLRV%: standardized future liver remnant volume ratio; FRLF: future remnant liver function; FRLV: future remnant liver volume; MUR: Mebrofenin uptake rate; FLR-C: percentage of counts within the future remnant liver; HIBA-i: the HIBA-index (the proportion of radionuclide accumulated in the future remnant liver); FLR/sTLV: the ratio between future liver remnant volume and standardized total liver volume; FLR/mTLV: the ratio between future liver remnant volume and measured total liver volume; FLR/BW: the ratio between future liver remnant volume and body weight; rLUV_(BSA): liver uptake value of the remnant liver corrected for body surface area; rLUR: remnant liver uptake ratio; LUV_(BW): liver uptake value of the remnant liver corrected for body weight; %remnant LF: the relative residual liver function; % remnant LV: the relative residual liver volume; rLUV_(LV): liver uptake value of the remnant liver corrected for liver volume; and TLV: total liver volume
For the diagnostic characteristics, sensitivity, specificity, negative predictive value, positive predictive value, and area under the curve were reported if available
^aTLF and HBS^BSA correspond to the same global liver function estimate from [^99mTc]Tc-mebrofenin hepatobiliary scintigraphy, but in the two papers they are named differently
^bAfter PVE
^cFor “signs of postoperative liver failure”
^dReported here for patients with a %FLR (volume-based) of 35–60%, further details are provided for patients > 60% FLR in the original paper
^eIn the whole patient population, for information on characteristics of NMI for patients with bilirubin level < 50 μmol/L at the time of the hepatobiliary scintigraphy, see the paper
^fNon-preserved versus preserved liver function on postoperative day 5; non-preserved referring to moderate–severe hepatic dysfunction on postoperative day 5; and preserved referring to no or mild hepatic dysfunction on postoperative day 5

Back to article page