Alpha-fetoprotein useful marker in HCC

Changes in serum alpha-fetoprotein levels over time correlated with the development of hepatocellular carcinoma in hepatitis C patients, wrote Dr. Elliot Lee and his colleagues in the April 1 issue of Clinical Gastroenterology and Hepatology.*

"If confirmed in future studies, these findings could be used to develop individualized risk assessments for clinical use, which could then influence the frequency and type of further testing" in this population, added the researchers (doi:10.1016/j.cgh.2012.11.029).

In what he called "the first large study to demonstrate that patterns of AFP [alpha-fetoprotein] over time are independently associated with HCC [hepatocellular carcinoma] development," Dr. Lee of the University of Michigan, Ann Arbor, looked at patients enrolled in the HALT-C (Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis) trial, in which hepatitis C patients who were nonresponders to prior antiviral therapy were randomized to receive maintenance pegylated interferon or placebo.

According to the trial protocol, patients were screened every 3 months for the first 3.5 years, then every 6 months thereafter on a voluntary basis, with levels of serum alpha-fetoprotein measured at each visit.

Patients who developed hepatocellular carcinoma were matched to controls without HCC in a 1:3 ratio according to the length of follow-up "in order to exclude bias caused by subjects with longer follow-up time having higher cumulative probability of HCC development, as well as more AFP values available," wrote the authors.

Overall, among 967 subjects with hepatitis C, 82 developed HCC during the study period, with a median of 18 (range, 5-22) AFP tests performed.

Dr. Lee then analyzed the association between the development of HCC and three AFP patterns.

First, the researchers assessed the rate of rise, defined as the patient’s final alpha-fetoprotein level (before study conclusion or HCC development, whichever came first) minus the AFP baseline value (on study entry), all over the follow-up duration divided by 90 days.

He found that this metric was associated with an odds ratio for developing HCC of 1.178 (P less than .001) in a simple logistic regression analysis that accounted for baseline risk factors only, with an area under the receiver-operating characteristic (AUROC) of 0.69.

Next, Dr. Lee calculated the standard deviation of AFP, defined as the standard deviation of all AFP values recorded within each patient.

In the same simple logistic regression, the standard deviation calculation was associated with an odds ratio of 1.026 for HCC per unit increase in standard deviation (P less than .001) and an AUROC of 0.76.

Third, Dr. Lee looked at the most recent AFP value on record. In this case, the odds ratio for developing HCC was 1.012 (also with P less than .001) and an AUROC of 0.76.

Finally, the researchers incorporated both standard deviation and rate of rise of AFP along with baseline age, platelet count, and smoking history to create a "history" model, which had an AUROC of 0.81 – an even better predictor than the models that looked only at one of those factors.

In an attempt to explain their findings, the researchers wrote, "it is intuitively evident why the rate of rise of AFP might be associated with risk of HCC development, but what might explain the association with standard deviation?"

They added, "The biologic basis is unknown, but we theorize that fluctuations in AFP may reflect cycles of damage and regeneration within the liver, and that growth factors involved in regeneration could stimulate hepatocarcinogenesis."

The authors acknowledged that the study had several limitations. "From a practical perspective, these metrics will only be useful once a patient has been followed for at least 2 years in order for the patterns to emerge," they wrote.

Additionally, "this study included only patients with hepatitis C; it is unknown whether these associations would be present among patients with other chronic liver diseases."

The authors said they had no relevant financial disclosures. They disclosed grant support from the National Science Foundation Graduate Research Fellowship.

*Correction, 3/25/2013: An earlier version of this story misstated the name of Clinical Gastroenterology and Hepatology.

Changes in serum alpha-fetoprotein levels over time correlated with the development of hepatocellular carcinoma in hepatitis C patients, wrote Dr. Elliot Lee and his colleagues in the April 1 issue of Clinical Gastroenterology and Hepatology.*

"If confirmed in future studies, these findings could be used to develop individualized risk assessments for clinical use, which could then influence the frequency and type of further testing" in this population, added the researchers (doi:10.1016/j.cgh.2012.11.029).

In what he called "the first large study to demonstrate that patterns of AFP [alpha-fetoprotein] over time are independently associated with HCC [hepatocellular carcinoma] development," Dr. Lee of the University of Michigan, Ann Arbor, looked at patients enrolled in the HALT-C (Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis) trial, in which hepatitis C patients who were nonresponders to prior antiviral therapy were randomized to receive maintenance pegylated interferon or placebo.

According to the trial protocol, patients were screened every 3 months for the first 3.5 years, then every 6 months thereafter on a voluntary basis, with levels of serum alpha-fetoprotein measured at each visit.

Patients who developed hepatocellular carcinoma were matched to controls without HCC in a 1:3 ratio according to the length of follow-up "in order to exclude bias caused by subjects with longer follow-up time having higher cumulative probability of HCC development, as well as more AFP values available," wrote the authors.

Overall, among 967 subjects with hepatitis C, 82 developed HCC during the study period, with a median of 18 (range, 5-22) AFP tests performed.

Dr. Lee then analyzed the association between the development of HCC and three AFP patterns.

First, the researchers assessed the rate of rise, defined as the patient’s final alpha-fetoprotein level (before study conclusion or HCC development, whichever came first) minus the AFP baseline value (on study entry), all over the follow-up duration divided by 90 days.

He found that this metric was associated with an odds ratio for developing HCC of 1.178 (P less than .001) in a simple logistic regression analysis that accounted for baseline risk factors only, with an area under the receiver-operating characteristic (AUROC) of 0.69.

Next, Dr. Lee calculated the standard deviation of AFP, defined as the standard deviation of all AFP values recorded within each patient.

In the same simple logistic regression, the standard deviation calculation was associated with an odds ratio of 1.026 for HCC per unit increase in standard deviation (P less than .001) and an AUROC of 0.76.

Third, Dr. Lee looked at the most recent AFP value on record. In this case, the odds ratio for developing HCC was 1.012 (also with P less than .001) and an AUROC of 0.76.

Finally, the researchers incorporated both standard deviation and rate of rise of AFP along with baseline age, platelet count, and smoking history to create a "history" model, which had an AUROC of 0.81 – an even better predictor than the models that looked only at one of those factors.

In an attempt to explain their findings, the researchers wrote, "it is intuitively evident why the rate of rise of AFP might be associated with risk of HCC development, but what might explain the association with standard deviation?"

They added, "The biologic basis is unknown, but we theorize that fluctuations in AFP may reflect cycles of damage and regeneration within the liver, and that growth factors involved in regeneration could stimulate hepatocarcinogenesis."

The authors acknowledged that the study had several limitations. "From a practical perspective, these metrics will only be useful once a patient has been followed for at least 2 years in order for the patterns to emerge," they wrote.

Additionally, "this study included only patients with hepatitis C; it is unknown whether these associations would be present among patients with other chronic liver diseases."

The authors said they had no relevant financial disclosures. They disclosed grant support from the National Science Foundation Graduate Research Fellowship.

*Correction, 3/25/2013: An earlier version of this story misstated the name of Clinical Gastroenterology and Hepatology.

Changes in serum alpha-fetoprotein levels over time correlated with the development of hepatocellular carcinoma in hepatitis C patients, wrote Dr. Elliot Lee and his colleagues in the April 1 issue of Clinical Gastroenterology and Hepatology.*

"If confirmed in future studies, these findings could be used to develop individualized risk assessments for clinical use, which could then influence the frequency and type of further testing" in this population, added the researchers (doi:10.1016/j.cgh.2012.11.029).

In what he called "the first large study to demonstrate that patterns of AFP [alpha-fetoprotein] over time are independently associated with HCC [hepatocellular carcinoma] development," Dr. Lee of the University of Michigan, Ann Arbor, looked at patients enrolled in the HALT-C (Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis) trial, in which hepatitis C patients who were nonresponders to prior antiviral therapy were randomized to receive maintenance pegylated interferon or placebo.

According to the trial protocol, patients were screened every 3 months for the first 3.5 years, then every 6 months thereafter on a voluntary basis, with levels of serum alpha-fetoprotein measured at each visit.

Patients who developed hepatocellular carcinoma were matched to controls without HCC in a 1:3 ratio according to the length of follow-up "in order to exclude bias caused by subjects with longer follow-up time having higher cumulative probability of HCC development, as well as more AFP values available," wrote the authors.

Overall, among 967 subjects with hepatitis C, 82 developed HCC during the study period, with a median of 18 (range, 5-22) AFP tests performed.

Dr. Lee then analyzed the association between the development of HCC and three AFP patterns.

First, the researchers assessed the rate of rise, defined as the patient’s final alpha-fetoprotein level (before study conclusion or HCC development, whichever came first) minus the AFP baseline value (on study entry), all over the follow-up duration divided by 90 days.

He found that this metric was associated with an odds ratio for developing HCC of 1.178 (P less than .001) in a simple logistic regression analysis that accounted for baseline risk factors only, with an area under the receiver-operating characteristic (AUROC) of 0.69.

Next, Dr. Lee calculated the standard deviation of AFP, defined as the standard deviation of all AFP values recorded within each patient.

In the same simple logistic regression, the standard deviation calculation was associated with an odds ratio of 1.026 for HCC per unit increase in standard deviation (P less than .001) and an AUROC of 0.76.

Third, Dr. Lee looked at the most recent AFP value on record. In this case, the odds ratio for developing HCC was 1.012 (also with P less than .001) and an AUROC of 0.76.

Finally, the researchers incorporated both standard deviation and rate of rise of AFP along with baseline age, platelet count, and smoking history to create a "history" model, which had an AUROC of 0.81 – an even better predictor than the models that looked only at one of those factors.

In an attempt to explain their findings, the researchers wrote, "it is intuitively evident why the rate of rise of AFP might be associated with risk of HCC development, but what might explain the association with standard deviation?"

They added, "The biologic basis is unknown, but we theorize that fluctuations in AFP may reflect cycles of damage and regeneration within the liver, and that growth factors involved in regeneration could stimulate hepatocarcinogenesis."

The authors acknowledged that the study had several limitations. "From a practical perspective, these metrics will only be useful once a patient has been followed for at least 2 years in order for the patterns to emerge," they wrote.

Additionally, "this study included only patients with hepatitis C; it is unknown whether these associations would be present among patients with other chronic liver diseases."

The authors said they had no relevant financial disclosures. They disclosed grant support from the National Science Foundation Graduate Research Fellowship.

*Correction, 3/25/2013: An earlier version of this story misstated the name of Clinical Gastroenterology and Hepatology.