All studies published in Clinical Gastroenterology and Hepatology are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.
Clinical Gastroenterology and Hepatology 2014;-:-–-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
Factors That Affect Efficacy of Ultrasound Surveillance for Early Stage Hepatocellular Carcinoma in Patients With Cirrhosis Q42
Paolo Del Poggio,* Stefano Olmi,* Francesca Ciccarese,* Mariella Di Marco,‡ Gian Ludovico Rapaccini,§ Luisa Benvegnù,k Franco Borzio,¶ Fabio Farinati,# Marco Zoli,** Edoardo Giovanni Giannini,‡‡ Eugenio Caturelli,§§ Maria Chiaramonte,kk and Franco Trevisani,¶¶ for the Italian Liver Cancer (ITA.LI.CA) Group *Unità di Epatologia, Policlinico S. Marco, Zingonia, Bergamo, Italy; ‡Divisione di Medicina, Azienda Ospedaliera Bolognini, Seriate, Italy; §Medicina Interna e Gastroenterologia, Università Cattolica, Rome, Italy; kMedicina Clinica e Sperimentale, # Scienze Chirurgiche e Gastroenterologiche, Università di Padova, Padova, Italy; ¶Medicina Interna ed Epatologia, Ospedale Fatebenefratelli, Milano, Italy; **Dipartimento di Scienze Mediche e Chirugiche, Medicina Interna, ¶¶Dipartimento di Scienze Mediche e Chirugiche, Semeiotica Medica, Alma Mater Studiorum–Università di Bologna, Bologna, Italy; ‡‡Gastroenterologia, Università di Genova, Genova, Italy; §§Gastroenterologia, Ospedale Belcolle, Viterbo, Italy; kkGastroenterologia, Ospedale Negrar, Verona, Italy BACKGROUND & AIMS:
Ultrasound surveillance does not detect early stage hepatocellular carcinomas (HCCs) in some patients with cirrhosis, although the reasons for this have not been well studied. We assessed the rate at which ultrasound fails to detect early stage HCCs and factors that affect its performance.
METHODS:
We collected information on 1170 consecutive patients included in the Italian Liver Cancer (ITA.LI.CA) database who had Child–Pugh A or B cirrhosis and were diagnosed with HCC during semiannual or annual ultrasound surveillance, from January 1987 through December 2008. Etiologies included hepatitis C virus infection (59.3%), alcohol abuse (11.3%), hepatitis B virus infection (9%), a combination of factors (15.6%), and other factors (4.7%). Surveillance was considered to be a failure when patients were diagnosed with HCC at a stage beyond the Milan criteria (1 nodule £5 cm or £3 nodules each £3 cm).
RESULTS:
Ultrasound surveillance failed to detect HCC in 34.3% of patients and more often in the annual program than in the semiannual one (41.3% vs 32.2%; P < .01). Nearly half of surveillance failures were associated with at least one indicator of aggressive HCC (levels of AFP >1000 ng/mL, infiltrating tumors, or vascular invasion and metastases). Semiannual surveillance, female sex, Child–Pugh class A, and a-fetoprotein levels of 200 ng/mL or less were associated independently with successful ultrasound screening for HCC.
CONCLUSIONS:
Based on our analysis of surveillance for HCC in patients with cirrhosis, the efficacy of ultrasound-based screening is acceptable. Ultrasound was least effective in identifying aggressive HCC, and at surveillance intervals of more than 6 months.
Q7
Q8
Q9
Keywords: Liver Cancer; Early Detection; Fibrosis; Survival.
Q10 Q11
epatocellular carcinoma (HCC) is the fifth most common tumor worldwide and the main cause of mortality of cirrhotic patients.1 Despite continuous therapeutic advances its prognosis remains poor because the majority of these tumors are identified at a late stage. Several cohort and one randomized control study have shown that regular ultrasound surveillance in high-risk populations can achieve this goal and increase survival.2–6 For this reason, all international guidelines recommend implementation of regular ultrasound surveillance in these patients.1,7
H
In the surveillance setting, ultrasound has been found to have an excellent specificity (>90%), but low sensitivity.3,8 Namely, a meta-analysis of 13 studies performed
Abbreviations used in this paper: AFP, a-fetoprotein; BMI, body mass index; CI, confidence interval; CT, computed tomography; HALT-C, ___________________; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; OR, odds ratio. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.02.025
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
2
117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
Q12
Q13
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
in tertiary care centers indicated that ultrasound sensitivity in detecting early stage HCC may be as low as 63%, and that the addition of a-fetoprotein (AFP) measurement adds marginal benefit.9 The majority of these studies, however, were performed in the 1990s, when ultrasound examinations were performed with old technology equipment. After the year 2000, newer probes with higher-density crystals, together with compound and harmonic imaging, have improved the chance of identifying small liver nodules. Ultrasound nonetheless had several limitations: first, it fails to detect infiltrative tumors; second, its sensitivity remarkably decreases in cirrhotic livers with coarse echo texture, in obese patients, in those with abdominal gas, or those not compliant with the breath-hold command; third, it is highly operatordependent and requires expertise, so that ultrasound examinations performed in real-life settings may have a lower yield.10 An analysis of the causes of surveillance failure to detect early HCC could lead to a better use of the ultrasound technique or highlight the need for alternative methods of surveillance in some cases. The aims of our study were to assess the failure rate of regular ultrasound surveillance in detecting an early HCC in a large population of cirrhotic patients recruited at several medical institutions, including both tertiary and primary care hospitals, and to identify the factors responsible for ultrasound failure.
Patients and Methods Patients
Q14
We retrospectively analyzed the data from the ITA.LI.CA database, including 3027 consecutive patients diagnosed with HCC from January 1987 through December 2008 at 11 Italian medical institutions (6 academic tertiary care centers and 5 hepatology/gastroenterology units located in general hospitals). The data were collected prospectively and updated every 2 years. Among the initial 3027 patients, we selected the 1170 Child–Pugh A or B cirrhotic patients diagnosed with HCC during semiannual or annual ultrasound surveillance. Child–Pugh C patients were excluded because practical guidelines do not recommend surveillance for these patients.1 AFP determinations were available only at the time of HCC diagnosis. The macroscopic HCC features of all Child–Pugh A and B patients of the ITA.LI.CA database are described in Supplementary Table 1. Cirrhosis was diagnosed histologically in 446 cases and by clinical, endoscopic, or ultrasound evaluation in the remaining cases. The criteria used for the classification of the etiology of cirrhosis, diagnosis, and staging of HCC are reported in the Supplementary Materials and Methods section. HCC was classified as biologically aggressive if, at the time of diagnosis, it showed at least one of the following: AFP
-,
No.
-
level greater than 1000 ng/mL, diffuse involvement of the liver on imaging, vascular invasion, or distant metastases. To be included in the study patients had to be diagnosed with HCC during ultrasound surveillance. We conventionally accepted that this was shown by the availability of an ultrasound examination negative for HCC, and performed in the 12 months preceding HCC diagnosis. If the last negative ultrasound was performed more than 1 year before the examination suggesting the HCC presence, the patient was excluded from the analysis because his/her surveillance was considered inconsistent. According to the surveillance interval, patients were divided into 2 groups, as follows. Group A contained patients with an ultrasound examination performed between 1 and 6 months before tumor diagnosis. Because most of the examinations were performed between 4 and 6 months before tumor diagnosis, we considered this group to have semiannual surveillance. Group B contained patients with an ultrasound examination performed between 7 and 12 months before HCC diagnosis. Because most of the examinations were performed between 7 and 12 months before tumor diagnosis, we considered this group to have annual surveillance. Surveillance failure was defined as follows: (1) the detection of a tumor beyond the Milan criteria after HCC staging had been completed; (2) the detection of HCC after a negative ultrasound by computed tomography (CT) or magnetic resonance imaging (MRI), which were prompted by insufficient ultrasound quality or an increase of AFP level to greater than 50 ng/mL. Surveillance failure was determined in 2 calendar periods (1987–1999 and 2000–2008), with the cut-off point set at the end of the 20th century because, at that time, the new compound and harmonic imaging were introduced in the ultrasound equipment.
Q15
Q16
Q17
Q18
Statistical Analysis Continuous data are expressed as medians and ranges or means standard deviations in case of nonnormal or normal data distribution, respectively. The Mann–Whitney U test was used to compare continuous data, and chi-square analysis or the Fisher exact test were used to compare discrete variables. The Pearson product moment was used to study correlations between continuous variables. Factors available in at least 90% of patients at the time of HCC diagnosis and potentially affecting ultrasound surveillance were investigated by univariate analysis. Factors associated with surveillance failure with a P value of .10 or less were tested by multiple logistic regression analysis with stepwise backward elimination of nonsignificant variables. The odds ratio (OR) and 95% confidence interval (95% CI) were reported. The Med Calc statistical package was used for these analyses. For
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q19
Q20
Q21
175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
-
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
2014
Ultrasound for HCC and Cirrhosis
further details on statistical analysis see Supplementary Materials and Methods section.
the
Results Of the 1430 cirrhotic patients undergoing ultrasound surveillance, 1170 patients eventually were available for the study (Figure 1). Their demographic and clinical characteristics are summarized in Table 1. HCC was diagnosed by noninvasive criteria in 1013 patients and histologically in 157 patients. In the latter cases, all studied variables were obtained at the time of biopsy, which was performed no later than 1 to 2 months after imaging. Surveillance was performed with ultrasound plus AFP determination in 1102 patients (94%) and with ultrasound alone in the remaining 68 patients (6%). Of the 1170 patients, 889 patients (76%) underwent semiannual surveillance with a mean interval of 5.3 1.29 months (range, 1–6 mo) and 281 patients (24%) underwent annual surveillance with a mean interval of 11.4 1.3 months (range, 7–12 mo) (Figure 1). In 80.5% of cases from the semiannual surveillance group, ultrasound examinations were performed every 4 to 6 months and in 19.5% every 1 to 3 months. In 86.3% of cases from the annual surveillance group examinations were performed every 10 to 12 months and in 13.7% they were performed every 7 to 9 months. Only 28 patients (10%) included in the latter group underwent
3
Table 1. Demographic and Clinical Characteristics of 1170 Cirrhotic Patients Diagnosed With HCC During Surveillance at Eleven ITA.LI.CA Centers (1987–2008) Median age, y (range) Sex M F Etiology of cirrhosis Hepatitis B virus Hepatitis C virus Alcohol Combined etiology Others Comorbid illness Yes No Histologic diagnosis Child–Pugh A B Median AFP level (range) Period of diagnosis 1987–1999 2000–2008 Diagnosis occurrence Tertiary centers Primary centers
Figure 1. Flow chart.
830 (70.9%) 340 (29.1%) 106 694 133 182 55
(9%) (59.3%) (11.3%) (15.6%) (4.7%)
710 (60.6%) 460 (39.4%) 468 (40%) 855 (73.1%) 315 (26.9%) 17 ng/mL (2–300,000) 438 (37.4%) 732 (62.6%) Q34
677 (58%) 493 (42%)
examinations with a 7- to 8-month periodicity. Overlap between the 2 surveillance groups therefore was excluded. In 97 patients (8.3% of the entire population) ultrasound was negative for HCC and the tumor was detected only by CT/MRI. In these cases, CT/MRI was performed because of the coarse echo texture of the liver (53 cases) or an AFP level greater than 50 ng/mL (44 cases). In 33 of these 97 cases (32%) HCC was diagnosed beyond Milan criteria, similar to what was observed in our entire population (32%). All of these 97 cases, in which HCC was detected only by CT/MRI, were considered as surveillance failures, even when HCC was diagnosed within Milan criteria (38 cases).
In the overall population, ultrasound surveillance detected HCC meeting Milan criteria in 949 patients (81%), but after they were staged with CT/MRI only 768 (65.6%) remained in this category; therefore, 402 cases were diagnosed beyond these criteria, yielding a surveillance failure rate of 34.3% after complete staging. Among the 768 patients meeting Milan criteria, only 235 (20% of the whole population) had a single tumor 2 cm or smaller, with most of them (207) being diagnosed during semiannual surveillance, so that their prevalence was 10% in the annual and 23.3% in the semiannual program (P < .001) Of these patients with tiny tumors, only 189 (16.1% of the whole population) could be
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q33
68 (29–92)
Surveillance Failure
Q32
Q1
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
4
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
Q22
Q23
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
included in BCLC stage 0 (very early), because 40 were in Child–Pugh B class and 6 had a performance status greater than 0. The prevalence of surveillance failure was distributed differently according to the surveillance interval, occurring in 286 (32.2%) patients undergoing semiannual surveillance and in 116 patients (41.3%) undergoing an annual program (P < .01). Notably, 113 cases (39.5%) of surveillance failures in the semiannual group and 79 (68.1%) in the annual group were associated with biologically aggressive tumors, as previously defined (Figure 1). The failure rate decreased over time, from 36.4% in the period from 1987 to 1999 to 33.0% in the period from 2000 to 2008, but the difference did not reach statistical significance (P ¼ .255) (Supplementary Figure 1). This trend was owing to both an increased number of early HCCs detected with the semiannual program and increased use of this program (from 64% to 83%; P < .001) over time. Surveillance failures differed among the ITA.LI.CA centers (ranging from 22% to 70% for 1987–1999, and from 10% to 64% for 2000–2008). The etiology of cirrhosis did not affect surveillance failure. In fact, HCC was detected at a late stage in 38 of 106 (36%) hepatitis B surface antigen–positive patients, in 239 of 694 (34.4%) anti–hepatitis C virus–positive patients, in 52 of 133 (39%) alcohol abusing patients, and in 71 of 182 (39%) combined etiology patients (P ¼ .572). Conversely, surveillance failure was associated with increasing AFP levels (P ¼ .02, Pearson product moment). Namely, compared with patients with an AFP level less than 10 ng/mL, the OR for surveillance failure was 1.36 (95% CI, 0.94–1.9) for an AFP level of 20 ng/mL or greater, 2.17 (95% CI, 1.58–2.9) for an AFP level of 50 ng/mL or greater, 2.69 (95% CI, 1.91–3.7) for an AFP level of 100 ng/mL or greater, and 3.12 (95% CI, 2.11–4.6) when the AFP level was 200 ng/mL or greater. However, AFP level scarcely was useful to predict surveillance failure: the positive predictive value for levels greater than 200 ng/mL was only 54%, whereas the negative predictive value of 100 ng/mL was only 69%. By using the upper normal value as a cut-off level (ie, 10 ng/mL), the negative predictive value increased to 72%. Even restricting the analysis to the 728 patients with ALT levels lower than 2 times the upper normal value, the diagnostic accuracy of AFP in predicting surveillance failure was low, with positive and negative predictive values never exceeding 72%. At univariate analysis, sex, surveillance interval, Child–Pugh class, and AFP level (dichotomized at 200 ng/mL) were found to be correlated with surveillance failure (Table 2). These variables and age were entered into the multiple logistic regression model that yielded annual surveillance (OR, 1.49; 95% CI, 1.11–1.99), male sex (OR, 1.41; 95% CI, 1.06–1.87), Child–Pugh class B (OR, 1.34, 95% CI, 1.01–1.7), and AFP level of 200 ng/mL or greater (OR, 2.6; 95% CI, 1.83–3.73) as factors independently associated with surveillance failure.
-,
No.
-
Table 2. Variables Associated With Surveillance Failure at Univariate and Multivariate Analysis Variable
Univariate analysis
Age: >68 (565) vs 68 y (605) Sex: M (826) vs F (344)
P ¼ .092b P ¼ .029b
ALT level: >2 (490) vs 2 ULN (680) Comorbidities: yes: (710) vs no (460) Type of center: tertiary (824) vs primary (346) Surveillance: annual (281) vs semiannual (889) Period of diagnosis: 2000–2008 (714) vs 1987–1999 (456) Child–Pugh class: B (315) vs A (855) Alcohol intake: M >30 g/d and F >20 g/d (639) vs others (531) Esophageal varices: yes (722) vs no (448) Platelet count (1150) AFP level: >200 (153) vs 200 ng/mL (949) Albumin level (1100)
P ¼ .724b
Multivariate analysisa 1.41c (1.06–1.87), P ¼ .0175
P ¼ .752b P ¼ .388b P ¼ .0063b P ¼ .219b P ¼ .030b P ¼ .729b
1.49c (1.11–1.99), P ¼ .0075
1.34c (1.01–1.78), P ¼ .0387
P ¼ .699b P ¼ .179d P ¼ .000b P ¼ .48d
2.6c (1.83–3.73), P < .0001
NOTE. In the first column the number of patients is shown in parentheses. Variables entered in multivariate analysis are shown in bold. ALT, alanine aminotransferase; ULN, upper limit of normal. a Multiple logistic regression analysis (backward). b Chi-square test. c Odds ratio (95% CI). d Pearson product moment correlation.
Discussion In our study, nearly one third of HCCs diagnosed during semiannual or annual ultrasound surveillance were detected beyond Milan criteria, which define the upper limit to apply curative treatments. As previously reported,11 surveillance failures were reduced significantly by the more stringent surveillance program, recommended by international guidelines. Notably, half of the failures observed in semiannually surveyed patients were associated with tumors presenting with features of biological aggressiveness, such as diffuse liver infiltration, vascular thrombosis, distant metastases, or AFP levels greater than 1000 ng/mL. In these cases, surveillance failure hardly could be ascribed to limitations of the periodic screening tool, and it is unlikely that a more stringent program could have produced an earlier diagnosis and changed the prognosis. This assumption indeed is supported by the results of a multicenter randomized prospective trial showing that in cirrhotic patients a trimestral surveillance was not superior to the standard 6-month program.12 Thus, it is established that the 6-month periodicity of ultrasound examination should neither be prolonged,13 nor shortened.12
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q35
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
-
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
Q24
Q25
2014
Ultrasound for HCC and Cirrhosis
The performance of semiannual ultrasound surveillance in our cirrhotic patients yielded fairly good results, particularly if we consider as true failures only those cases (19%) in whom the procedure failed the early detection of nonaggressive tumors. Similar results have been reported by the aforementioned French study, in which 25% of HCCs found during an every 3- or 6-month surveillance were beyond the Milan criteria, but only 9% lacked features of aggressiveness and were unequivocally imputable to intrinsic limitations of the ultrasound technique.12 However, this a posteriori interpretation of the relationship between surveillance failure and tumor aggressiveness may be biased by the fact that HCC can acquire an aggressive phenotype with time, so that its detection at a late stage may allow even slow-growing tumors to achieve features of aggressiveness. Indeed, in our patients the prevalence of surveillance failures associated with aggressive tumors increased from 39.5% in the semiannual program to 68.1% in the annual program. This risk warns against the prolongation of the surveillance interval beyond 6 months. It nevertheless is important to underline that single HCCs 2 cm or less were detected in only 20% of cases, similarly to the HALT-C trial population,14 and mainly during semiannual monitoring. Therefore, it can be concluded that the 6-month interval is preferable, even if in real-world clinical practice an optimal result is achieved in only a minority of patients. According to a number of studies,13,15,16 we confirmed an increased accuracy of ultrasound surveillance over time. The surveillance outcome could be improved further by the identification of patients at high risk of failure, in whom a different screening modality might be proposed.17 Pertinently, we found that male sex, Child–Pugh B class, and AFP levels were associated independently with failures. Nonetheless, caution should be applied in considering AFP as a suitable predictor of surveillance failure because we could not identify an efficient threshold with acceptable diagnostic accuracy for this purpose, even using values greater than 200 ng/mL and after excluding patients with high necroinflammatory activity, which may affect AFP levels.18 It could be assumed that the association between AFP and surveillance failure was spurious, being related to tumor aggressiveness, a variable not included in the multivariate analysis because it was not available at the time of surveillance implementation. It also has been suggested that the incorporation of the trend over time of AFP level in surveillance may decrease surveillance failures and improve the detection rate of early stage HCCs,19 but we could not test this assumption because our database does not contain the AFP values antecedent to the HCC diagnosis. The relationship between male sex and surveillance failure is enigmatic. A higher body mass index (BMI) and an increased prevalence of steatosis in men, making the visualization of the liver more difficult, presumably were causative factors. We could not assess the role of BMI
5
because these data were reported in only one fifth of our records. However, by performing a multivariate analysis in the 267 patients in whom anthropometric measures were available, we found that only annual surveillance and overweight (BMI > 25 kg/mg) were associated significantly with surveillance failure (Supplementary Table 2). This suggests that a high body weight rather than male sex predicts failure. Our findings were different from what was found in the HALT-C trial population, in which surveillance failure was not influenced by BMI,14 although it should be pointed out that the HALT analysis was underpowered because of the small number of HCCs detected beyond the early stage. The association between Child–Pugh class B and surveillance failure may be attributed intuitively to the increasing prevalence of a coarse liver echo texture as the disease progresses, hampering the identification of nodules. To conclude in a pragmatic way, overweight and advanced liver disease (coarse liver pattern) are seemingly the key factors in ultrasound surveillance failure. The use of CT scanning or MRI in these “difficult” patients is probably not cost effective for their high cost, false-positive rate,8,20 and, for CT, the radiation risk, although the utility of contrast-enhanced ultrasonography still remains to be explored. Intravascular contrast agents (Levovist and SonoVue) did not prove to be useful in the detection of small HCCs owing to the impracticality of scanning the entire liver during the arterial phase, when small tumors are enhanced.21 On the other hand, perfluorocarbon (Sonazoid) seems promising because this agent is taken up and retained by Kupffer cells for 60 minutes after bolus injection, allowing a scan of the entire liver in the late phase when HCC appears as a black hole.22 Sonazoid-enhanced ultrasound therefore might overcome the adverse impact of a coarse echo texture, but not that of obesity or severe fatty livers, in which CT or MRI remain the only option to accurately exclude HCC. New biomarkers or the improved use of AFP, such as incorporating the trend over time of AFP level in ultrasound surveillance of patients at high risk of surveillance failure could improve the detection rate of early HCC.19,23 Another issue that should be addressed in considering surveillance failure is the expertise of sonographers. In ITA.LI.CA centers, ultrasound examinations are performed by dedicated ultrasonographers, but in a number of our patients it was performed by general radiologists outside our centers. The lack of information on the proportion of patients periodically screened at or outside each ITA.LI.CA centers did not allow us to compare the performance of specialized and general sonographers, and to analyze possible differences between primary and tertiary referral centers. Another point we could not analyze was the influence of patient’s adherence to surveillance, which was shown as an important determinant of surveillance failures in some studies,10,24,25 but not in the HALT-C trial,14 in
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q26
Q27 Q28
Q29 Q30
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
6
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638
Q31
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
which the absence of detection and not lack of adherence was the strongest determinant of surveillance failures. Finally, we conventionally considered patients under surveillance as only those who underwent at least one ultrasound examination in the year preceding cancer detection but, because of the retrospective nature of the study, we could not distinguish with certainty if the last negative ultrasound examination was performed as a surveillance test or for other clinical reasons. In conclusion, our retrospective study shows that in clinical practice, assuming that patients are compliant, the semiannual program of ultrasound surveillance is superior to the annual program, and the risk of true surveillance failure is acceptably low. This failure is related chiefly to tumor aggressiveness, but also to patients being overweight and coarse hepatic echo texture, limiting the detection of small HCCs. To improve the cost effectiveness of surveillance it is important to select the patients appropriately in whom to apply additional screening modalities,26 accordingly to both their HCC risk27 and to the risk of ultrasound surveillance failure. Despite this, a certain number of failures will continue to affect the effectiveness of this practice because its causes, except being overweight, are not manageable, and CT or MRI cannot be claimed as suitable substitutes of ultrasound.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2014.02.025.
References 1. EASL-EORTC. Clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56:908–943. 2. Wong LL, Limm WM, Severino R, et al. Improved survival with screening for hepatocellular carcinoma. Liver Transpl 2000; 6:320–325. 3. El-Serag HB, Kramer JR, Chen GJ, et al. Effectiveness of AFP and ultrasound tests on hepatocellular carcinoma mortality in HCV-infected patients in the USA. Gut 2011;60:992–997. 4. Pateron D, Ganne N, Trinchet JC, et al. Prospective study of screening of hepatocellular carcinoma in Caucasian patients with cirrhosis. J Hepatol 1994;20:65–71. 5. Trevisani F, De Notariis S, Rapaccini G, et al. Semiannual and annual surveillance of cirrhotic patients for hepatocellular carcinoma: effects on cancer stage and survival. Am J Gastroenterol 2002;97:734–744. 6. Zhang BH, Yang BH, Tang ZY. Randomized control trial of screening for hepatocellular carcinoma. J Cancer Res Clin Oncol 2004;130:417–422. 7. Bruix J, Sherman M. American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011;53:1020–1022. 8. Singal AG, Conjeevaram HS, Volk ML, et al. Effectiveness of hepatocellular carcinoma surveillance in patients with cirrhosis. Cancer Epidemiol Biomarkers Prev 2012;21:793–799.
-,
No.
-
9. Singal AG, Volk ML, Waljee A, et al. Meta-analysis: surveillance with ultrasound for early-stage hepatocellular carcinoma in patients with cirrhosis. Aliment Pharmacol Ther 2009;30: 37–47. 10. Davila JA, Morgan RO, Richardson PA, et al. Use of surveillance for hepatocellular carcinoma among patients with cirrhosis in the United States. Hepatology 2010;52:132–141. 11. Santi V, Buccione D, Di Micoli A, et al. The changing scenario of hepatocellular carcinoma over the last two decades in Italy. J Hepatol 2012;56:397–405. 12. Trinchet JC, Chaffaut C, Bourcier V, et al. Ultrasonographic surveillance of hepatocellular carcinoma in cirrhosis: a randomized trial comparing 3 and 6 month periodicity. Hepatology 2011;54:1987–1997. 13. Santi V, Trevisani F, Gramenzi A, et al. Semiannual surveillance is superior to annual surveillance for the detection of early hepatocellular carcinoma and patient survival. J Hepatol 2010;53: 291–297. 14. Singal AG, Nehra M, Adams-Huet B, et al. Detection of hepatocellular carcinoma at advanced stages among patients in the HALT-C trial: where does surveillance fail? Am J Gastroenterol 2013;108:425–432. 15. Noda I, Kitamoto M, Nakahara H, et al. Regular surveillance by imaging for early detection and better prognosis of hepatocellular carcinoma in patients infected with hepatitis C virus. J Gastroenterol 2010;45:105–112. 16. Kuo YH, Lu SN, Chen CL, et al. Hepatocellular carcinoma surveillance and appropriate treatment options improve survival for patients with cirrhosis. Eur J Cancer 2010;46: 744–751. 17. Lee JM, Trevisani F, Vilgrain V, et al. Imaging diagnosis and staging of hepatocellular carcinoma. Liver Tranplant 2011;17: S34–S43. 18. Richardson P, Duran Z, Kramer J. Determinants of serum alphafetoprotein levels in hepatitis C infected patients. Clin Gastroenterol Hepatol 2012;10:428–433. 19. Lee E, Edward S, Singal A, et al. Improving screening for hepatocellular carcinoma by incorporating data on levels of alphafetoprotein over time. Clin Gastroenterol Hepatol 2013;11: 437–440. 20. Lencioni R, Crocetti L, Della Pina MC, et al. Guidelines for imaging focal lesions in liver cirrhosis. Exp Rev Gastroenterol Hepatol 2008;2:697–703. 21. Lencioni R, Piscaglia F, Bolondi L. Contrast-enhanced ultrasound in the diagnosis of hepatocellular carcinoma. J Hepatol 2008;48:848–857. 22. Kudo M. Double contrast ultrasound: a novel surveillance tool for hepatocellular carcinoma. Am J Gastroenterol 2011;106: 368–370. 23. El Serag H, Kanwal F. Alpha-fetoprotein in hepatocellular carcinoma surveillance: mend it but do not end it. Clin Gastroenterol Hepatol 2013;11:441–443. 24. Davila JA, Henderson L, Kramer JR, et al. Utilization of surveillance for hepatocellular carcinoma among hepatitis C virusinfected veterans in the United States. Ann Intern Med 2011; 154:85–93. 25. Singal AG, Yopp A, Skinner CS, et al. Utilization of hepatocellular carcinoma surveillance among American patients: a systematic review. J Gen Intern Med 2012;27:861–867. 26. El-Serag HB. Surveillance for hepatocellular carcinoma: long way to achieve effectiveness. Dig Dis Sci 2012;57:3050–3051.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
-
2014
Ultrasound for HCC and Cirrhosis
27. Cucchetti A, Trevisani F, Cascon M, et al. Cost-effectiveness of 697 semi-annual surveillance for hepatocellular carcinoma in 698 cirrhotic patients of the Italian liver cancer population. J Hepatol 699 2012;56:1089–1096. 700 701 702 Reprint requests requests for reprints to: Paolo Del Poggio, MD, Unità di Epatologia, 703 Q2 Q3 Address Policlinico San Marco, Zingonia, Corso Europa Unita 7, 24040 Zingonia-Osio 704 Sotto, Bergamo, Italy. e-mail: [email protected]; fax: (39) 0363352889. 705 706 Q4 Acknowledgments 707 Q5 Other ITA.LI.CA members are as follows: Dipartimento di Scienze Mediche e Chirurgiche, Alma Mater Studiorum–Università di Bologna: Mauro Bernardi, 708 Maurizio Biselli, Paolo Caraceni, Alessandro Cucchetti, Marco Domenicali, 709 Marta Frigerio, Annagiulia Gramenzi, Francesca Garuti, Barbara Lenzi, Don710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
7
atella Magalotti, and Matteo Ravaioli; Dipartimento di Medicina Clinica e Sperimentale, Università di Padova: Alfredo Alberti, Angelo Gatta, and Maurizio Gios; Dipartimento di Scienze Chirurgiche e Gastroenterologiche, Università di Padova: Anna Giacomin, Veronica Vanin, Caterina Pozzan, and Gemma Maddalo; Dipartimento di Malattie Apparato Digerente e Medicina Interna, Azienda Ospedaliero–Universitaria di Bologna, Unità Operativa di Radiologia: Alberta Cappelli, Emanuela Giampalma, Rita Golfieri, Cristina Mosconi, and Matteo Renzulli; Unità Operativa di Gastroenterologia, Ospedale Belcolle, Viterbo: Paola Roselli; Unità Operativa di Gastroenterologia, Unità Operativa di Malattie Infettive in Ambito Penitenziario, Ospedale Belcolle, Viterbo: Paola Roselli, Serena Dell’Isola, and Anna Maria Ialungo; Dipartimento di Medicina Interna, Unità di Gastroenterologia, Università di Genova: Domenico Risso, Giorgio Sammito, Simona Marenco, and Linda Bruzzone; Unità Operativa di Medicina Interna e Gastroenterologia, Università Cattolica di Roma: Giulia Bosco.
Conflicts of interest The authors disclose no conflicts.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q6
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
7.e1
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
Q36
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
Supplementary Materials and Methods Criteria Used to Define the Etiology of Cirrhosis, Diagnosis, and Staging of Hepatocellular Carcinoma
Q37
Q38
Etiology of cirrhosis. The cause of liver disease was classified as follows: hepatitis C virus if patients were positive for serum anti–hepatitis C virus antibodies; hepatitis B virus if patients were surface antigen–positive carriers; alcoholic if the daily alcohol consumption was more than 60 g for women and more than 80 g for men for more than 10 years, in the absence of any other known causes of liver disease; combined if the disease had 2 or more of the earlier-described causes; and as other (cryptogenic liver disease, hereditary hemochromatosis, primary liver cirrhosis, nonalcoholic liver disease). Diagnosis and staging of hepatocellular carcinoma. The diagnosis of HCC was based on histology or by imaging according to the different EASL/AASLD guidelines published overtime. HCC was staged with an imaging work-up, including multiphasic abdominal CT or MRI and chest radiograph or CT, according to the judgment of the referring clinician. Additional diagnostic procedures were performed when clinically appropriate. For the purpose of this study, HCC was staged as follows: HCC meeting the Milan criteria (1 nodule 5 cm or 2–3 nodules 3 cm each, without vascular invasion and extrahepatic spread), and beyond the Milan criteria.
-,
No.
-
Additional Statistical Data Factors available in at least 90% of patients at the time of HCC diagnosis and potentially affecting ultrasound surveillance (age, sex, alanine aminotransferase, etiology, inappropriate alcohol consumption [>30 g/d for men and >20 g/d for women], Child–Pugh class, esophageal varices, platelet count, AFP level, albumin level, comorbidities [cardiovascular, pulmonary, renal, gastrointestinal and hematologic diseases, obesity, and diabetes], surveillance interval, type of center [primary vs tertiary referral center], and time of diagnosis [1987–1999 vs 2000–2008]) were investigated by univariate analysis. Factors associated with surveillance failure with a P value of .10 or less at univariate analysis were tested by multiple logistic regression analysis with stepwise backward elimination of nonsignificant variables. Continuous variables, if not otherwise specified, were dichotomized according to their median values. The odds ratio and 95% confidence intervals were reported. Multiple colinearity was checked by calculating the variance inflation factor and excluding variables with a variance inflation factor greater than 5. The variance inflation factor values of the included variables were 1.01 for semiannual surveillance, 1.04 for platelet count, 1.2 for Child–Pugh classification, and 1.16 for esophageal varices. A small degree of multicollinearity therefore was present, but never exceeded the value of 5 and did not affect the analysis.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928
-
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991
2014
Ultrasound for HCC and Cirrhosis
7.e2
Supplementary Figure 1. Surveillance failures according to the time of diagnosis (1987–1999 vs 2000–2008). White bars indicate the total number of HCCs diagnosed in each period during annual and semiannual surveillance. Black bars indicate the number of surveillance failures in the 2 programs (annual, 66 [41.5%] in the first period and 50 [40.6%] in the second period [P ¼ .98 NS; chi-square test]; semiannual, 100 [33.6%] and 186 [31.4%], respectively (P ¼ .55)]. The prevalence of surveillance failures decreased over time, although not significantly, from 36.4% in the period from 1987 to 1999 to 33.0% in the period from 2000 to 2008 (P ¼ .255 NS; chi-square test).
Q39
Supplementary Table 2. Variables Associated With Surveillance Failure in the 267 Patients With BMI Available Variable
Supplementary Table 1. Tumor Features of the Child–Pugh A/B Patients Included in the ITA.LI.CA Database
Tumor burden Within Milan criteria Single < 2 cm Single 2–5 cm 3 nodules 3 cm Outside Milan criteria
Q40
HCC diagnosed HCC diagnosed under ultrasound outside ultrasound surveillance (1170)a surveillance (1248) 833 256 445 132 337
(71.2%) (21.9%) (38.0%) (11.3%) (28.8%)
475 86 197 192 773
(38.1%) (6.9%) (15.8%) (15.4%) (61.9%)
NOTE. Relevant data were available for 2418 of 2468 patients: (1675 Child– Pugh A and 793 Child–Pugh B). The reported tumor characteristics were those obtained after an imaging study including not only ultrasonography but also CT scan or MRI. At the time we conducted this analysis the database had been updated until 2008 because a temporary interruption of scheduled updating was in place to allow the generation of a new online data collection system (http://www.progettoitalica.it) and the transfer of previously collected data. a Ultrasound examinations were performed not only by hepatologists or radiologists of the ITA.LI.CA centers, with specific skills in liver ultrasound, but also by general radiologists in outside facilities, according to the distance between the patient’s home and the ITA.LI.CA center, and his/her ability to regularly come to the referral center.
Univariate analysis
Age: >68 (150) vs 68 y (117) Sex: M (186) vs F (81) ALT level: >2 ULN (85) vs £2 ULN (182) Comorbidities: yes: (192) vs no (75) BMI: 25 (158) vs >25 (109)
P ¼ .675b P ¼ .174b P ¼ .034b
Type of center: tertiary (81) vs general (186) Surveillance: annual (57) vs semiannual (210) Child–Pugh class: B (55) vs A (212) Alcohol intake: Male >30 g/d and female >20 g/d (148) vs others (119) Esophageal varices: yes (109) vs no (158) Platelet count (267) AFP (264) Albumin (260)
P ¼ .297b
Multivariate analysisa
P ¼ .347b P ¼ .034b
P ¼ .027b P ¼ .536b
1.15 (1.04–1.25),d P < .01 1.14 (1.01–1.28),d P ¼ .02
P ¼ .741b P ¼ .631b P ¼ .038c P ¼ .050c P ¼ .051c
NOTE. In the first column the number of patients is shown in parentheses. Variables entered in the multivariate analysis are shown in bold. a Multiple logistic regression analysis (backward). b Chi-square. c Pearson product moment correlation. d Odds ratio (95% CI).
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q41
992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
Clinical Gastroenterology and Hepatology 2014;-:-–-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
Factors That Affect Efficacy of Ultrasound Surveillance for Early Stage Hepatocellular Carcinoma in Patients With Cirrhosis Q42
Paolo Del Poggio,* Stefano Olmi,* Francesca Ciccarese,* Mariella Di Marco,‡ Gian Ludovico Rapaccini,§ Luisa Benvegnù,k Franco Borzio,¶ Fabio Farinati,# Marco Zoli,** Edoardo Giovanni Giannini,‡‡ Eugenio Caturelli,§§ Maria Chiaramonte,kk and Franco Trevisani,¶¶ for the Italian Liver Cancer (ITA.LI.CA) Group *Unità di Epatologia, Policlinico S. Marco, Zingonia, Bergamo, Italy; ‡Divisione di Medicina, Azienda Ospedaliera Bolognini, Seriate, Italy; §Medicina Interna e Gastroenterologia, Università Cattolica, Rome, Italy; kMedicina Clinica e Sperimentale, # Scienze Chirurgiche e Gastroenterologiche, Università di Padova, Padova, Italy; ¶Medicina Interna ed Epatologia, Ospedale Fatebenefratelli, Milano, Italy; **Dipartimento di Scienze Mediche e Chirugiche, Medicina Interna, ¶¶Dipartimento di Scienze Mediche e Chirugiche, Semeiotica Medica, Alma Mater Studiorum–Università di Bologna, Bologna, Italy; ‡‡Gastroenterologia, Università di Genova, Genova, Italy; §§Gastroenterologia, Ospedale Belcolle, Viterbo, Italy; kkGastroenterologia, Ospedale Negrar, Verona, Italy BACKGROUND & AIMS:
Ultrasound surveillance does not detect early stage hepatocellular carcinomas (HCCs) in some patients with cirrhosis, although the reasons for this have not been well studied. We assessed the rate at which ultrasound fails to detect early stage HCCs and factors that affect its performance.
METHODS:
We collected information on 1170 consecutive patients included in the Italian Liver Cancer (ITA.LI.CA) database who had Child–Pugh A or B cirrhosis and were diagnosed with HCC during semiannual or annual ultrasound surveillance, from January 1987 through December 2008. Etiologies included hepatitis C virus infection (59.3%), alcohol abuse (11.3%), hepatitis B virus infection (9%), a combination of factors (15.6%), and other factors (4.7%). Surveillance was considered to be a failure when patients were diagnosed with HCC at a stage beyond the Milan criteria (1 nodule £5 cm or £3 nodules each £3 cm).
RESULTS:
Ultrasound surveillance failed to detect HCC in 34.3% of patients and more often in the annual program than in the semiannual one (41.3% vs 32.2%; P < .01). Nearly half of surveillance failures were associated with at least one indicator of aggressive HCC (levels of AFP >1000 ng/mL, infiltrating tumors, or vascular invasion and metastases). Semiannual surveillance, female sex, Child–Pugh class A, and a-fetoprotein levels of 200 ng/mL or less were associated independently with successful ultrasound screening for HCC.
CONCLUSIONS:
Based on our analysis of surveillance for HCC in patients with cirrhosis, the efficacy of ultrasound-based screening is acceptable. Ultrasound was least effective in identifying aggressive HCC, and at surveillance intervals of more than 6 months.
Q7
Q8
Q9
Keywords: Liver Cancer; Early Detection; Fibrosis; Survival.
Q10 Q11
epatocellular carcinoma (HCC) is the fifth most common tumor worldwide and the main cause of mortality of cirrhotic patients.1 Despite continuous therapeutic advances its prognosis remains poor because the majority of these tumors are identified at a late stage. Several cohort and one randomized control study have shown that regular ultrasound surveillance in high-risk populations can achieve this goal and increase survival.2–6 For this reason, all international guidelines recommend implementation of regular ultrasound surveillance in these patients.1,7
H
In the surveillance setting, ultrasound has been found to have an excellent specificity (>90%), but low sensitivity.3,8 Namely, a meta-analysis of 13 studies performed
Abbreviations used in this paper: AFP, a-fetoprotein; BMI, body mass index; CI, confidence interval; CT, computed tomography; HALT-C, ___________________; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; OR, odds ratio. © 2014 by the AGA Institute 1542-3565/$36.00 http://dx.doi.org/10.1016/j.cgh.2014.02.025
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116
2
117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
Q12
Q13
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
in tertiary care centers indicated that ultrasound sensitivity in detecting early stage HCC may be as low as 63%, and that the addition of a-fetoprotein (AFP) measurement adds marginal benefit.9 The majority of these studies, however, were performed in the 1990s, when ultrasound examinations were performed with old technology equipment. After the year 2000, newer probes with higher-density crystals, together with compound and harmonic imaging, have improved the chance of identifying small liver nodules. Ultrasound nonetheless had several limitations: first, it fails to detect infiltrative tumors; second, its sensitivity remarkably decreases in cirrhotic livers with coarse echo texture, in obese patients, in those with abdominal gas, or those not compliant with the breath-hold command; third, it is highly operatordependent and requires expertise, so that ultrasound examinations performed in real-life settings may have a lower yield.10 An analysis of the causes of surveillance failure to detect early HCC could lead to a better use of the ultrasound technique or highlight the need for alternative methods of surveillance in some cases. The aims of our study were to assess the failure rate of regular ultrasound surveillance in detecting an early HCC in a large population of cirrhotic patients recruited at several medical institutions, including both tertiary and primary care hospitals, and to identify the factors responsible for ultrasound failure.
Patients and Methods Patients
Q14
We retrospectively analyzed the data from the ITA.LI.CA database, including 3027 consecutive patients diagnosed with HCC from January 1987 through December 2008 at 11 Italian medical institutions (6 academic tertiary care centers and 5 hepatology/gastroenterology units located in general hospitals). The data were collected prospectively and updated every 2 years. Among the initial 3027 patients, we selected the 1170 Child–Pugh A or B cirrhotic patients diagnosed with HCC during semiannual or annual ultrasound surveillance. Child–Pugh C patients were excluded because practical guidelines do not recommend surveillance for these patients.1 AFP determinations were available only at the time of HCC diagnosis. The macroscopic HCC features of all Child–Pugh A and B patients of the ITA.LI.CA database are described in Supplementary Table 1. Cirrhosis was diagnosed histologically in 446 cases and by clinical, endoscopic, or ultrasound evaluation in the remaining cases. The criteria used for the classification of the etiology of cirrhosis, diagnosis, and staging of HCC are reported in the Supplementary Materials and Methods section. HCC was classified as biologically aggressive if, at the time of diagnosis, it showed at least one of the following: AFP
-,
No.
-
level greater than 1000 ng/mL, diffuse involvement of the liver on imaging, vascular invasion, or distant metastases. To be included in the study patients had to be diagnosed with HCC during ultrasound surveillance. We conventionally accepted that this was shown by the availability of an ultrasound examination negative for HCC, and performed in the 12 months preceding HCC diagnosis. If the last negative ultrasound was performed more than 1 year before the examination suggesting the HCC presence, the patient was excluded from the analysis because his/her surveillance was considered inconsistent. According to the surveillance interval, patients were divided into 2 groups, as follows. Group A contained patients with an ultrasound examination performed between 1 and 6 months before tumor diagnosis. Because most of the examinations were performed between 4 and 6 months before tumor diagnosis, we considered this group to have semiannual surveillance. Group B contained patients with an ultrasound examination performed between 7 and 12 months before HCC diagnosis. Because most of the examinations were performed between 7 and 12 months before tumor diagnosis, we considered this group to have annual surveillance. Surveillance failure was defined as follows: (1) the detection of a tumor beyond the Milan criteria after HCC staging had been completed; (2) the detection of HCC after a negative ultrasound by computed tomography (CT) or magnetic resonance imaging (MRI), which were prompted by insufficient ultrasound quality or an increase of AFP level to greater than 50 ng/mL. Surveillance failure was determined in 2 calendar periods (1987–1999 and 2000–2008), with the cut-off point set at the end of the 20th century because, at that time, the new compound and harmonic imaging were introduced in the ultrasound equipment.
Q15
Q16
Q17
Q18
Statistical Analysis Continuous data are expressed as medians and ranges or means standard deviations in case of nonnormal or normal data distribution, respectively. The Mann–Whitney U test was used to compare continuous data, and chi-square analysis or the Fisher exact test were used to compare discrete variables. The Pearson product moment was used to study correlations between continuous variables. Factors available in at least 90% of patients at the time of HCC diagnosis and potentially affecting ultrasound surveillance were investigated by univariate analysis. Factors associated with surveillance failure with a P value of .10 or less were tested by multiple logistic regression analysis with stepwise backward elimination of nonsignificant variables. The odds ratio (OR) and 95% confidence interval (95% CI) were reported. The Med Calc statistical package was used for these analyses. For
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q19
Q20
Q21
175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
-
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
2014
Ultrasound for HCC and Cirrhosis
further details on statistical analysis see Supplementary Materials and Methods section.
the
Results Of the 1430 cirrhotic patients undergoing ultrasound surveillance, 1170 patients eventually were available for the study (Figure 1). Their demographic and clinical characteristics are summarized in Table 1. HCC was diagnosed by noninvasive criteria in 1013 patients and histologically in 157 patients. In the latter cases, all studied variables were obtained at the time of biopsy, which was performed no later than 1 to 2 months after imaging. Surveillance was performed with ultrasound plus AFP determination in 1102 patients (94%) and with ultrasound alone in the remaining 68 patients (6%). Of the 1170 patients, 889 patients (76%) underwent semiannual surveillance with a mean interval of 5.3 1.29 months (range, 1–6 mo) and 281 patients (24%) underwent annual surveillance with a mean interval of 11.4 1.3 months (range, 7–12 mo) (Figure 1). In 80.5% of cases from the semiannual surveillance group, ultrasound examinations were performed every 4 to 6 months and in 19.5% every 1 to 3 months. In 86.3% of cases from the annual surveillance group examinations were performed every 10 to 12 months and in 13.7% they were performed every 7 to 9 months. Only 28 patients (10%) included in the latter group underwent
3
Table 1. Demographic and Clinical Characteristics of 1170 Cirrhotic Patients Diagnosed With HCC During Surveillance at Eleven ITA.LI.CA Centers (1987–2008) Median age, y (range) Sex M F Etiology of cirrhosis Hepatitis B virus Hepatitis C virus Alcohol Combined etiology Others Comorbid illness Yes No Histologic diagnosis Child–Pugh A B Median AFP level (range) Period of diagnosis 1987–1999 2000–2008 Diagnosis occurrence Tertiary centers Primary centers
Figure 1. Flow chart.
830 (70.9%) 340 (29.1%) 106 694 133 182 55
(9%) (59.3%) (11.3%) (15.6%) (4.7%)
710 (60.6%) 460 (39.4%) 468 (40%) 855 (73.1%) 315 (26.9%) 17 ng/mL (2–300,000) 438 (37.4%) 732 (62.6%) Q34
677 (58%) 493 (42%)
examinations with a 7- to 8-month periodicity. Overlap between the 2 surveillance groups therefore was excluded. In 97 patients (8.3% of the entire population) ultrasound was negative for HCC and the tumor was detected only by CT/MRI. In these cases, CT/MRI was performed because of the coarse echo texture of the liver (53 cases) or an AFP level greater than 50 ng/mL (44 cases). In 33 of these 97 cases (32%) HCC was diagnosed beyond Milan criteria, similar to what was observed in our entire population (32%). All of these 97 cases, in which HCC was detected only by CT/MRI, were considered as surveillance failures, even when HCC was diagnosed within Milan criteria (38 cases).
In the overall population, ultrasound surveillance detected HCC meeting Milan criteria in 949 patients (81%), but after they were staged with CT/MRI only 768 (65.6%) remained in this category; therefore, 402 cases were diagnosed beyond these criteria, yielding a surveillance failure rate of 34.3% after complete staging. Among the 768 patients meeting Milan criteria, only 235 (20% of the whole population) had a single tumor 2 cm or smaller, with most of them (207) being diagnosed during semiannual surveillance, so that their prevalence was 10% in the annual and 23.3% in the semiannual program (P < .001) Of these patients with tiny tumors, only 189 (16.1% of the whole population) could be
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q33
68 (29–92)
Surveillance Failure
Q32
Q1
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
4
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
Q22
Q23
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
included in BCLC stage 0 (very early), because 40 were in Child–Pugh B class and 6 had a performance status greater than 0. The prevalence of surveillance failure was distributed differently according to the surveillance interval, occurring in 286 (32.2%) patients undergoing semiannual surveillance and in 116 patients (41.3%) undergoing an annual program (P < .01). Notably, 113 cases (39.5%) of surveillance failures in the semiannual group and 79 (68.1%) in the annual group were associated with biologically aggressive tumors, as previously defined (Figure 1). The failure rate decreased over time, from 36.4% in the period from 1987 to 1999 to 33.0% in the period from 2000 to 2008, but the difference did not reach statistical significance (P ¼ .255) (Supplementary Figure 1). This trend was owing to both an increased number of early HCCs detected with the semiannual program and increased use of this program (from 64% to 83%; P < .001) over time. Surveillance failures differed among the ITA.LI.CA centers (ranging from 22% to 70% for 1987–1999, and from 10% to 64% for 2000–2008). The etiology of cirrhosis did not affect surveillance failure. In fact, HCC was detected at a late stage in 38 of 106 (36%) hepatitis B surface antigen–positive patients, in 239 of 694 (34.4%) anti–hepatitis C virus–positive patients, in 52 of 133 (39%) alcohol abusing patients, and in 71 of 182 (39%) combined etiology patients (P ¼ .572). Conversely, surveillance failure was associated with increasing AFP levels (P ¼ .02, Pearson product moment). Namely, compared with patients with an AFP level less than 10 ng/mL, the OR for surveillance failure was 1.36 (95% CI, 0.94–1.9) for an AFP level of 20 ng/mL or greater, 2.17 (95% CI, 1.58–2.9) for an AFP level of 50 ng/mL or greater, 2.69 (95% CI, 1.91–3.7) for an AFP level of 100 ng/mL or greater, and 3.12 (95% CI, 2.11–4.6) when the AFP level was 200 ng/mL or greater. However, AFP level scarcely was useful to predict surveillance failure: the positive predictive value for levels greater than 200 ng/mL was only 54%, whereas the negative predictive value of 100 ng/mL was only 69%. By using the upper normal value as a cut-off level (ie, 10 ng/mL), the negative predictive value increased to 72%. Even restricting the analysis to the 728 patients with ALT levels lower than 2 times the upper normal value, the diagnostic accuracy of AFP in predicting surveillance failure was low, with positive and negative predictive values never exceeding 72%. At univariate analysis, sex, surveillance interval, Child–Pugh class, and AFP level (dichotomized at 200 ng/mL) were found to be correlated with surveillance failure (Table 2). These variables and age were entered into the multiple logistic regression model that yielded annual surveillance (OR, 1.49; 95% CI, 1.11–1.99), male sex (OR, 1.41; 95% CI, 1.06–1.87), Child–Pugh class B (OR, 1.34, 95% CI, 1.01–1.7), and AFP level of 200 ng/mL or greater (OR, 2.6; 95% CI, 1.83–3.73) as factors independently associated with surveillance failure.
-,
No.
-
Table 2. Variables Associated With Surveillance Failure at Univariate and Multivariate Analysis Variable
Univariate analysis
Age: >68 (565) vs 68 y (605) Sex: M (826) vs F (344)
P ¼ .092b P ¼ .029b
ALT level: >2 (490) vs 2 ULN (680) Comorbidities: yes: (710) vs no (460) Type of center: tertiary (824) vs primary (346) Surveillance: annual (281) vs semiannual (889) Period of diagnosis: 2000–2008 (714) vs 1987–1999 (456) Child–Pugh class: B (315) vs A (855) Alcohol intake: M >30 g/d and F >20 g/d (639) vs others (531) Esophageal varices: yes (722) vs no (448) Platelet count (1150) AFP level: >200 (153) vs 200 ng/mL (949) Albumin level (1100)
P ¼ .724b
Multivariate analysisa 1.41c (1.06–1.87), P ¼ .0175
P ¼ .752b P ¼ .388b P ¼ .0063b P ¼ .219b P ¼ .030b P ¼ .729b
1.49c (1.11–1.99), P ¼ .0075
1.34c (1.01–1.78), P ¼ .0387
P ¼ .699b P ¼ .179d P ¼ .000b P ¼ .48d
2.6c (1.83–3.73), P < .0001
NOTE. In the first column the number of patients is shown in parentheses. Variables entered in multivariate analysis are shown in bold. ALT, alanine aminotransferase; ULN, upper limit of normal. a Multiple logistic regression analysis (backward). b Chi-square test. c Odds ratio (95% CI). d Pearson product moment correlation.
Discussion In our study, nearly one third of HCCs diagnosed during semiannual or annual ultrasound surveillance were detected beyond Milan criteria, which define the upper limit to apply curative treatments. As previously reported,11 surveillance failures were reduced significantly by the more stringent surveillance program, recommended by international guidelines. Notably, half of the failures observed in semiannually surveyed patients were associated with tumors presenting with features of biological aggressiveness, such as diffuse liver infiltration, vascular thrombosis, distant metastases, or AFP levels greater than 1000 ng/mL. In these cases, surveillance failure hardly could be ascribed to limitations of the periodic screening tool, and it is unlikely that a more stringent program could have produced an earlier diagnosis and changed the prognosis. This assumption indeed is supported by the results of a multicenter randomized prospective trial showing that in cirrhotic patients a trimestral surveillance was not superior to the standard 6-month program.12 Thus, it is established that the 6-month periodicity of ultrasound examination should neither be prolonged,13 nor shortened.12
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q35
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464
-
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522
Q24
Q25
2014
Ultrasound for HCC and Cirrhosis
The performance of semiannual ultrasound surveillance in our cirrhotic patients yielded fairly good results, particularly if we consider as true failures only those cases (19%) in whom the procedure failed the early detection of nonaggressive tumors. Similar results have been reported by the aforementioned French study, in which 25% of HCCs found during an every 3- or 6-month surveillance were beyond the Milan criteria, but only 9% lacked features of aggressiveness and were unequivocally imputable to intrinsic limitations of the ultrasound technique.12 However, this a posteriori interpretation of the relationship between surveillance failure and tumor aggressiveness may be biased by the fact that HCC can acquire an aggressive phenotype with time, so that its detection at a late stage may allow even slow-growing tumors to achieve features of aggressiveness. Indeed, in our patients the prevalence of surveillance failures associated with aggressive tumors increased from 39.5% in the semiannual program to 68.1% in the annual program. This risk warns against the prolongation of the surveillance interval beyond 6 months. It nevertheless is important to underline that single HCCs 2 cm or less were detected in only 20% of cases, similarly to the HALT-C trial population,14 and mainly during semiannual monitoring. Therefore, it can be concluded that the 6-month interval is preferable, even if in real-world clinical practice an optimal result is achieved in only a minority of patients. According to a number of studies,13,15,16 we confirmed an increased accuracy of ultrasound surveillance over time. The surveillance outcome could be improved further by the identification of patients at high risk of failure, in whom a different screening modality might be proposed.17 Pertinently, we found that male sex, Child–Pugh B class, and AFP levels were associated independently with failures. Nonetheless, caution should be applied in considering AFP as a suitable predictor of surveillance failure because we could not identify an efficient threshold with acceptable diagnostic accuracy for this purpose, even using values greater than 200 ng/mL and after excluding patients with high necroinflammatory activity, which may affect AFP levels.18 It could be assumed that the association between AFP and surveillance failure was spurious, being related to tumor aggressiveness, a variable not included in the multivariate analysis because it was not available at the time of surveillance implementation. It also has been suggested that the incorporation of the trend over time of AFP level in surveillance may decrease surveillance failures and improve the detection rate of early stage HCCs,19 but we could not test this assumption because our database does not contain the AFP values antecedent to the HCC diagnosis. The relationship between male sex and surveillance failure is enigmatic. A higher body mass index (BMI) and an increased prevalence of steatosis in men, making the visualization of the liver more difficult, presumably were causative factors. We could not assess the role of BMI
5
because these data were reported in only one fifth of our records. However, by performing a multivariate analysis in the 267 patients in whom anthropometric measures were available, we found that only annual surveillance and overweight (BMI > 25 kg/mg) were associated significantly with surveillance failure (Supplementary Table 2). This suggests that a high body weight rather than male sex predicts failure. Our findings were different from what was found in the HALT-C trial population, in which surveillance failure was not influenced by BMI,14 although it should be pointed out that the HALT analysis was underpowered because of the small number of HCCs detected beyond the early stage. The association between Child–Pugh class B and surveillance failure may be attributed intuitively to the increasing prevalence of a coarse liver echo texture as the disease progresses, hampering the identification of nodules. To conclude in a pragmatic way, overweight and advanced liver disease (coarse liver pattern) are seemingly the key factors in ultrasound surveillance failure. The use of CT scanning or MRI in these “difficult” patients is probably not cost effective for their high cost, false-positive rate,8,20 and, for CT, the radiation risk, although the utility of contrast-enhanced ultrasonography still remains to be explored. Intravascular contrast agents (Levovist and SonoVue) did not prove to be useful in the detection of small HCCs owing to the impracticality of scanning the entire liver during the arterial phase, when small tumors are enhanced.21 On the other hand, perfluorocarbon (Sonazoid) seems promising because this agent is taken up and retained by Kupffer cells for 60 minutes after bolus injection, allowing a scan of the entire liver in the late phase when HCC appears as a black hole.22 Sonazoid-enhanced ultrasound therefore might overcome the adverse impact of a coarse echo texture, but not that of obesity or severe fatty livers, in which CT or MRI remain the only option to accurately exclude HCC. New biomarkers or the improved use of AFP, such as incorporating the trend over time of AFP level in ultrasound surveillance of patients at high risk of surveillance failure could improve the detection rate of early HCC.19,23 Another issue that should be addressed in considering surveillance failure is the expertise of sonographers. In ITA.LI.CA centers, ultrasound examinations are performed by dedicated ultrasonographers, but in a number of our patients it was performed by general radiologists outside our centers. The lack of information on the proportion of patients periodically screened at or outside each ITA.LI.CA centers did not allow us to compare the performance of specialized and general sonographers, and to analyze possible differences between primary and tertiary referral centers. Another point we could not analyze was the influence of patient’s adherence to surveillance, which was shown as an important determinant of surveillance failures in some studies,10,24,25 but not in the HALT-C trial,14 in
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q26
Q27 Q28
Q29 Q30
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
6
581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638
Q31
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
which the absence of detection and not lack of adherence was the strongest determinant of surveillance failures. Finally, we conventionally considered patients under surveillance as only those who underwent at least one ultrasound examination in the year preceding cancer detection but, because of the retrospective nature of the study, we could not distinguish with certainty if the last negative ultrasound examination was performed as a surveillance test or for other clinical reasons. In conclusion, our retrospective study shows that in clinical practice, assuming that patients are compliant, the semiannual program of ultrasound surveillance is superior to the annual program, and the risk of true surveillance failure is acceptably low. This failure is related chiefly to tumor aggressiveness, but also to patients being overweight and coarse hepatic echo texture, limiting the detection of small HCCs. To improve the cost effectiveness of surveillance it is important to select the patients appropriately in whom to apply additional screening modalities,26 accordingly to both their HCC risk27 and to the risk of ultrasound surveillance failure. Despite this, a certain number of failures will continue to affect the effectiveness of this practice because its causes, except being overweight, are not manageable, and CT or MRI cannot be claimed as suitable substitutes of ultrasound.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://dx.doi.org/10.1016/j.cgh.2014.02.025.
References 1. EASL-EORTC. Clinical practice guidelines: management of hepatocellular carcinoma. J Hepatol 2012;56:908–943. 2. Wong LL, Limm WM, Severino R, et al. Improved survival with screening for hepatocellular carcinoma. Liver Transpl 2000; 6:320–325. 3. El-Serag HB, Kramer JR, Chen GJ, et al. Effectiveness of AFP and ultrasound tests on hepatocellular carcinoma mortality in HCV-infected patients in the USA. Gut 2011;60:992–997. 4. Pateron D, Ganne N, Trinchet JC, et al. Prospective study of screening of hepatocellular carcinoma in Caucasian patients with cirrhosis. J Hepatol 1994;20:65–71. 5. Trevisani F, De Notariis S, Rapaccini G, et al. Semiannual and annual surveillance of cirrhotic patients for hepatocellular carcinoma: effects on cancer stage and survival. Am J Gastroenterol 2002;97:734–744. 6. Zhang BH, Yang BH, Tang ZY. Randomized control trial of screening for hepatocellular carcinoma. J Cancer Res Clin Oncol 2004;130:417–422. 7. Bruix J, Sherman M. American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma: an update. Hepatology 2011;53:1020–1022. 8. Singal AG, Conjeevaram HS, Volk ML, et al. Effectiveness of hepatocellular carcinoma surveillance in patients with cirrhosis. Cancer Epidemiol Biomarkers Prev 2012;21:793–799.
-,
No.
-
9. Singal AG, Volk ML, Waljee A, et al. Meta-analysis: surveillance with ultrasound for early-stage hepatocellular carcinoma in patients with cirrhosis. Aliment Pharmacol Ther 2009;30: 37–47. 10. Davila JA, Morgan RO, Richardson PA, et al. Use of surveillance for hepatocellular carcinoma among patients with cirrhosis in the United States. Hepatology 2010;52:132–141. 11. Santi V, Buccione D, Di Micoli A, et al. The changing scenario of hepatocellular carcinoma over the last two decades in Italy. J Hepatol 2012;56:397–405. 12. Trinchet JC, Chaffaut C, Bourcier V, et al. Ultrasonographic surveillance of hepatocellular carcinoma in cirrhosis: a randomized trial comparing 3 and 6 month periodicity. Hepatology 2011;54:1987–1997. 13. Santi V, Trevisani F, Gramenzi A, et al. Semiannual surveillance is superior to annual surveillance for the detection of early hepatocellular carcinoma and patient survival. J Hepatol 2010;53: 291–297. 14. Singal AG, Nehra M, Adams-Huet B, et al. Detection of hepatocellular carcinoma at advanced stages among patients in the HALT-C trial: where does surveillance fail? Am J Gastroenterol 2013;108:425–432. 15. Noda I, Kitamoto M, Nakahara H, et al. Regular surveillance by imaging for early detection and better prognosis of hepatocellular carcinoma in patients infected with hepatitis C virus. J Gastroenterol 2010;45:105–112. 16. Kuo YH, Lu SN, Chen CL, et al. Hepatocellular carcinoma surveillance and appropriate treatment options improve survival for patients with cirrhosis. Eur J Cancer 2010;46: 744–751. 17. Lee JM, Trevisani F, Vilgrain V, et al. Imaging diagnosis and staging of hepatocellular carcinoma. Liver Tranplant 2011;17: S34–S43. 18. Richardson P, Duran Z, Kramer J. Determinants of serum alphafetoprotein levels in hepatitis C infected patients. Clin Gastroenterol Hepatol 2012;10:428–433. 19. Lee E, Edward S, Singal A, et al. Improving screening for hepatocellular carcinoma by incorporating data on levels of alphafetoprotein over time. Clin Gastroenterol Hepatol 2013;11: 437–440. 20. Lencioni R, Crocetti L, Della Pina MC, et al. Guidelines for imaging focal lesions in liver cirrhosis. Exp Rev Gastroenterol Hepatol 2008;2:697–703. 21. Lencioni R, Piscaglia F, Bolondi L. Contrast-enhanced ultrasound in the diagnosis of hepatocellular carcinoma. J Hepatol 2008;48:848–857. 22. Kudo M. Double contrast ultrasound: a novel surveillance tool for hepatocellular carcinoma. Am J Gastroenterol 2011;106: 368–370. 23. El Serag H, Kanwal F. Alpha-fetoprotein in hepatocellular carcinoma surveillance: mend it but do not end it. Clin Gastroenterol Hepatol 2013;11:441–443. 24. Davila JA, Henderson L, Kramer JR, et al. Utilization of surveillance for hepatocellular carcinoma among hepatitis C virusinfected veterans in the United States. Ann Intern Med 2011; 154:85–93. 25. Singal AG, Yopp A, Skinner CS, et al. Utilization of hepatocellular carcinoma surveillance among American patients: a systematic review. J Gen Intern Med 2012;27:861–867. 26. El-Serag HB. Surveillance for hepatocellular carcinoma: long way to achieve effectiveness. Dig Dis Sci 2012;57:3050–3051.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
-
2014
Ultrasound for HCC and Cirrhosis
27. Cucchetti A, Trevisani F, Cascon M, et al. Cost-effectiveness of 697 semi-annual surveillance for hepatocellular carcinoma in 698 cirrhotic patients of the Italian liver cancer population. J Hepatol 699 2012;56:1089–1096. 700 701 702 Reprint requests requests for reprints to: Paolo Del Poggio, MD, Unità di Epatologia, 703 Q2 Q3 Address Policlinico San Marco, Zingonia, Corso Europa Unita 7, 24040 Zingonia-Osio 704 Sotto, Bergamo, Italy. e-mail: [email protected]; fax: (39) 0363352889. 705 706 Q4 Acknowledgments 707 Q5 Other ITA.LI.CA members are as follows: Dipartimento di Scienze Mediche e Chirurgiche, Alma Mater Studiorum–Università di Bologna: Mauro Bernardi, 708 Maurizio Biselli, Paolo Caraceni, Alessandro Cucchetti, Marco Domenicali, 709 Marta Frigerio, Annagiulia Gramenzi, Francesca Garuti, Barbara Lenzi, Don710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754
7
atella Magalotti, and Matteo Ravaioli; Dipartimento di Medicina Clinica e Sperimentale, Università di Padova: Alfredo Alberti, Angelo Gatta, and Maurizio Gios; Dipartimento di Scienze Chirurgiche e Gastroenterologiche, Università di Padova: Anna Giacomin, Veronica Vanin, Caterina Pozzan, and Gemma Maddalo; Dipartimento di Malattie Apparato Digerente e Medicina Interna, Azienda Ospedaliero–Universitaria di Bologna, Unità Operativa di Radiologia: Alberta Cappelli, Emanuela Giampalma, Rita Golfieri, Cristina Mosconi, and Matteo Renzulli; Unità Operativa di Gastroenterologia, Ospedale Belcolle, Viterbo: Paola Roselli; Unità Operativa di Gastroenterologia, Unità Operativa di Malattie Infettive in Ambito Penitenziario, Ospedale Belcolle, Viterbo: Paola Roselli, Serena Dell’Isola, and Anna Maria Ialungo; Dipartimento di Medicina Interna, Unità di Gastroenterologia, Università di Genova: Domenico Risso, Giorgio Sammito, Simona Marenco, and Linda Bruzzone; Unità Operativa di Medicina Interna e Gastroenterologia, Università Cattolica di Roma: Giulia Bosco.
Conflicts of interest The authors disclose no conflicts.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q6
755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812
7.e1
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
Q36
Del Poggio et al
Clinical Gastroenterology and Hepatology Vol.
Supplementary Materials and Methods Criteria Used to Define the Etiology of Cirrhosis, Diagnosis, and Staging of Hepatocellular Carcinoma
Q37
Q38
Etiology of cirrhosis. The cause of liver disease was classified as follows: hepatitis C virus if patients were positive for serum anti–hepatitis C virus antibodies; hepatitis B virus if patients were surface antigen–positive carriers; alcoholic if the daily alcohol consumption was more than 60 g for women and more than 80 g for men for more than 10 years, in the absence of any other known causes of liver disease; combined if the disease had 2 or more of the earlier-described causes; and as other (cryptogenic liver disease, hereditary hemochromatosis, primary liver cirrhosis, nonalcoholic liver disease). Diagnosis and staging of hepatocellular carcinoma. The diagnosis of HCC was based on histology or by imaging according to the different EASL/AASLD guidelines published overtime. HCC was staged with an imaging work-up, including multiphasic abdominal CT or MRI and chest radiograph or CT, according to the judgment of the referring clinician. Additional diagnostic procedures were performed when clinically appropriate. For the purpose of this study, HCC was staged as follows: HCC meeting the Milan criteria (1 nodule 5 cm or 2–3 nodules 3 cm each, without vascular invasion and extrahepatic spread), and beyond the Milan criteria.
-,
No.
-
Additional Statistical Data Factors available in at least 90% of patients at the time of HCC diagnosis and potentially affecting ultrasound surveillance (age, sex, alanine aminotransferase, etiology, inappropriate alcohol consumption [>30 g/d for men and >20 g/d for women], Child–Pugh class, esophageal varices, platelet count, AFP level, albumin level, comorbidities [cardiovascular, pulmonary, renal, gastrointestinal and hematologic diseases, obesity, and diabetes], surveillance interval, type of center [primary vs tertiary referral center], and time of diagnosis [1987–1999 vs 2000–2008]) were investigated by univariate analysis. Factors associated with surveillance failure with a P value of .10 or less at univariate analysis were tested by multiple logistic regression analysis with stepwise backward elimination of nonsignificant variables. Continuous variables, if not otherwise specified, were dichotomized according to their median values. The odds ratio and 95% confidence intervals were reported. Multiple colinearity was checked by calculating the variance inflation factor and excluding variables with a variance inflation factor greater than 5. The variance inflation factor values of the included variables were 1.01 for semiannual surveillance, 1.04 for platelet count, 1.2 for Child–Pugh classification, and 1.16 for esophageal varices. A small degree of multicollinearity therefore was present, but never exceeded the value of 5 and did not affect the analysis.
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928
-
929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991
2014
Ultrasound for HCC and Cirrhosis
7.e2
Supplementary Figure 1. Surveillance failures according to the time of diagnosis (1987–1999 vs 2000–2008). White bars indicate the total number of HCCs diagnosed in each period during annual and semiannual surveillance. Black bars indicate the number of surveillance failures in the 2 programs (annual, 66 [41.5%] in the first period and 50 [40.6%] in the second period [P ¼ .98 NS; chi-square test]; semiannual, 100 [33.6%] and 186 [31.4%], respectively (P ¼ .55)]. The prevalence of surveillance failures decreased over time, although not significantly, from 36.4% in the period from 1987 to 1999 to 33.0% in the period from 2000 to 2008 (P ¼ .255 NS; chi-square test).
Q39
Supplementary Table 2. Variables Associated With Surveillance Failure in the 267 Patients With BMI Available Variable
Supplementary Table 1. Tumor Features of the Child–Pugh A/B Patients Included in the ITA.LI.CA Database
Tumor burden Within Milan criteria Single < 2 cm Single 2–5 cm 3 nodules 3 cm Outside Milan criteria
Q40
HCC diagnosed HCC diagnosed under ultrasound outside ultrasound surveillance (1170)a surveillance (1248) 833 256 445 132 337
(71.2%) (21.9%) (38.0%) (11.3%) (28.8%)
475 86 197 192 773
(38.1%) (6.9%) (15.8%) (15.4%) (61.9%)
NOTE. Relevant data were available for 2418 of 2468 patients: (1675 Child– Pugh A and 793 Child–Pugh B). The reported tumor characteristics were those obtained after an imaging study including not only ultrasonography but also CT scan or MRI. At the time we conducted this analysis the database had been updated until 2008 because a temporary interruption of scheduled updating was in place to allow the generation of a new online data collection system (http://www.progettoitalica.it) and the transfer of previously collected data. a Ultrasound examinations were performed not only by hepatologists or radiologists of the ITA.LI.CA centers, with specific skills in liver ultrasound, but also by general radiologists in outside facilities, according to the distance between the patient’s home and the ITA.LI.CA center, and his/her ability to regularly come to the referral center.
Univariate analysis
Age: >68 (150) vs 68 y (117) Sex: M (186) vs F (81) ALT level: >2 ULN (85) vs £2 ULN (182) Comorbidities: yes: (192) vs no (75) BMI: 25 (158) vs >25 (109)
P ¼ .675b P ¼ .174b P ¼ .034b
Type of center: tertiary (81) vs general (186) Surveillance: annual (57) vs semiannual (210) Child–Pugh class: B (55) vs A (212) Alcohol intake: Male >30 g/d and female >20 g/d (148) vs others (119) Esophageal varices: yes (109) vs no (158) Platelet count (267) AFP (264) Albumin (260)
P ¼ .297b
Multivariate analysisa
P ¼ .347b P ¼ .034b
P ¼ .027b P ¼ .536b
1.15 (1.04–1.25),d P < .01 1.14 (1.01–1.28),d P ¼ .02
P ¼ .741b P ¼ .631b P ¼ .038c P ¼ .050c P ¼ .051c
NOTE. In the first column the number of patients is shown in parentheses. Variables entered in the multivariate analysis are shown in bold. a Multiple logistic regression analysis (backward). b Chi-square. c Pearson product moment correlation. d Odds ratio (95% CI).
FLA 5.2.0 DTD YJCGH53725_proof 17 May 2014 1:00 pm ce
Q41
992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
