Efficacy Evaluation of TheraSphere to Treat Inoperable Liver Cancer With Blockage of the Portal Vein: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy Evaluation of TheraSphere to Treat Inoperable Liver Cancer With Blockage of the Portal Vein (NCT01887717), a randomized phase 3 study. The primary outcome is Overall Survival (OS) From Time of Randomization.

2. Clinical Question

More broadly, the trial was designed to assess efficacy and safety of TheraSphere against sorafenib in participants with this form of advanced liver cancer. The primary outcome was overall survival, defined as time from randomization to death from any cause, with participants still alive censored at their last contact date. The registry also lists outcomes involving progression, portal vein thrombosis, tumor response, quality of life, and treatment-emergent adverse events. These are distinct questions and should not be treated as interchangeable measures of benefit or harm.

The clinical question is therefore a comparison between two assigned treatment strategies in a randomized, parallel-group trial. The posted outcome data describe survival medians among the stated treated population; they do not, on their own, establish whether the observed difference is statistically distinguishable from chance or attributable to treatment.

3. Trial Design

This was a prospective, randomized, active-controlled, parallel-group Phase 3 trial. It compared TheraSphere with sorafenib among participants with inoperable liver cancer and portal vein blockage. The registry describes the study as open-label, meaning participants and clinicians were not masked to assigned treatment. The sponsor was Boston Scientific Corporation.

The registry records enrollment of 36 participants. It gives a start date of 2014-02-27, a primary completion date of 2017-05-23, and a status of terminated. The reason for termination is not reported in the registry record. The primary outcome timeframe is randomization through death, with a stated maximum time of up to Month 30. That timeframe describes the outcome window, not necessarily the length of follow-up for every participant.

The TheraSphere arm received treatment through the hepatic artery to the treated lobe of the liver. The registry describes a target dose of 120 Gy + 10% and permits dose reduction to a minimum of 80 Gy + 10%. The sorafenib arm received oral tablets at 400 milligrams twice daily, with treatment continuing until clinical benefit ceased or toxicity became unacceptable; medically appropriate dose adjustments were permitted. These are the registry's treatment descriptions, not evidence that every participant received the same duration or dose.

The primary outcome analysis population is described as participants who received at least 1 dose of either treatment. Posted arm counts for that outcome are 15 for TheraSphere and 16 for sorafenib. The registry does not explain how these counts relate to the enrollment total or describe the disposition of all enrolled participants. It also does not report the randomization ratio, allocation concealment, eligibility criteria in detail, or the reason the trial ended.

The registry lists seven secondary outcomes: time to progression, time to worsening portal vein thrombosis, time to symptomatic progression, tumor response, quality-of-life change, time to quality-of-life deterioration, and treatment-emergent adverse events. The record reports no analysis count for the posted outcomes. The results therefore offer limited information for assessing the wider set of efficacy and safety questions beyond the primary survival endpoint.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants were randomized 1:1 to the TheraSphere group or control group. To balance the treatment groups, participants were stratified at randomization on the basis of serum level of alpha-fetoprotein (AFP) (\<400 nanograms/milliliter \[ng/mL\] versus \>400 ng/mL) and study center.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who received at least 1 dose of TheraSphere or Sorafenib.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Participants who received at least 1 dose of TheraSphere or Sorafenib.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS) From Time of Randomization, measured over Randomization up to participant's death (maximum time = up to Month 30).

The primary endpoint is overall survival from randomization. It measures the time from assignment to treatment until death from any cause. For participants who remained alive, the registry says their observation was censored at the date of last contact. Overall survival is commonly used in clinical trials because it is a direct endpoint: it does not depend on a judgment about whether a tumor has progressed, and death from any cause is the event of interest.

The registry reports the median overall survival in months for each arm. The median is the point at which the estimated survival distribution reaches its midpoint; it is not the average survival time, nor does it describe the experience of every participant. For TheraSphere, the reported median is 14.5 months, with a full range from 2.3 to 22.0 months. For sorafenib, the reported median is 18.2 months, with a full range from 2.9 to 18.2 months. A full range gives the observed endpoints of the reported values; it is not a confidence interval and does not quantify uncertainty around the median.

Survival analysis is useful when participants have different follow-up durations or remain alive when their observation ends. Censoring allows their known follow-up time to contribute without treating their last contact as a death. General statistical interpretation requires attention to how many participants remain at risk over time and how censoring is handled. Those details, along with uncertainty intervals for the medians, are not reported in the registry record.

The posted medians describe survival in the outcome population, but a comparison of medians alone is not a complete treatment-effect analysis. The registry does not report a hazard ratio, confidence interval, or test for the difference between treatment groups. Thus, the endpoint is clinically interpretable, while the precision and statistical strength of the between-arm comparison cannot be determined from the posted primary results alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a time-to-event outcome. The event is death from any cause, and the time origin is randomization. Participants who were alive at their last contact are censored at that date, as stated in the registry. This type of analysis can use information from participants whose event has not yet occurred, rather than requiring every participant to have died before analysis. The registry reports medians and full ranges by treatment arm.

What the median describes

General statistical explanation: the median survival time is the time at which the estimated proportion surviving reaches one-half. It is a summary of the survival distribution, not a prediction for an individual and not the mean time to death. The TheraSphere median is reported as 14.5 months, and the sorafenib median as 18.2 months. The corresponding reported full ranges are 2.3 to 22.0 months and 2.9 to 18.2 months.

A full range records the lowest and highest observed values in the reported summary. It can be strongly affected by values at the extremes and does not express the statistical uncertainty around a median. A confidence interval for each median would be a more direct way to describe its estimation uncertainty, but such intervals are not reported in the registry record. The record also does not provide survival probabilities at specified follow-up times or the number of participants still at risk over time.

How censoring enters the analysis

General statistical explanation: censoring occurs when a participant's event time is not observed during the available follow-up, but the participant's time under observation is known. For overall survival, an alive participant censored at last contact has known survival through that contact, while their later survival time is unknown. A survival method can incorporate that partial information. Its interpretation depends on assumptions about censoring and on sufficiently detailed follow-up data. The registry defines censoring at last contact for participants alive but does not describe further censoring rules, losses to follow-up, or the analysis procedure used to estimate the medians.

The registry's reported maximum outcome timeframe is up to Month 30. This does not mean that every participant had that length of follow-up. It also does not supply the timing or extent of censoring. Those details would help readers evaluate how much information supports later portions of the survival distribution.

What is needed for a comparative estimate

General statistical explanation: a treatment comparison for a time-to-event endpoint is often expressed using a hazard ratio, a confidence interval, and a statistical test. The hazard ratio summarizes a relative event rate under the chosen model, while the confidence interval reflects uncertainty in the estimate. The validity and meaning of that summary depend on the model and its assumptions. The registry record does not report a hazard ratio, confidence interval, p-value, or the statistical model used for the primary comparison. It therefore is not possible to evaluate those assumptions or quantify the comparative uncertainty from the posted results.

The medians can be read descriptively, but they do not substitute for the full survival curves or a model-based comparison. A median-based contrast also does not reveal whether any difference is concentrated early or late in follow-up. No numerical comparison should be constructed from the two medians as a replacement for an analysis that was not reported.

Randomization and the analyzed population

Random allocation is intended to balance prognostic factors across arms on average, supporting a fair comparison of assigned strategies. The registry identifies the trial as randomized and parallel-group, but does not report the allocation method, ratio, or concealment procedure. The primary population is described as participants who received at least 1 dose. An analysis of all randomized participants, often called an intention-to-treat analysis, retains people in their assigned groups regardless of treatment received. Whether such an analysis was conducted is not reported in the registry record.

There are 15 participants listed in the TheraSphere primary outcome arm and 16 in the sorafenib arm, alongside registry enrollment of 36. The record does not give a participant flow accounting for these figures. Without those details, it is not possible to determine how many randomized participants were excluded from this outcome population, why any were excluded, or how exclusions might affect the comparison.

Planning assumptions and absent analysis details

General statistical explanation: a sample-size calculation is a planning exercise based on a target effect, significance level, allocation, and model assumptions. It is not an estimate of the observed effect and does not describe the probability that a result is correct. The planning illustration associated with this analysis is based on the observed effect, uses a two-sided significance level of 0.05, assumes equal allocation, and uses a normal approximation. For time-to-event outcomes, the illustration is based on a total number of events. These are planning assumptions, not a description of the trial's original design. The trial's own sample-size justification and design assumptions are not reported in the registry record.

The registry does not report an interim-analysis plan, stopping boundaries, a prespecified statistical analysis plan, or a method for adjusting for multiple outcomes. These omissions limit assessment of whether the reported endpoint and analysis followed a prespecified plan. The status is terminated, but the record does not state why. In the absence of a reported comparison and analysis plan, the medians should be treated as descriptive results rather than as a complete statistical test of efficacy.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists the primary outcome as overall survival and lists seven secondary outcomes involving progression, portal vein thrombosis, symptomatic progression, tumor response, quality of life, and treatment-emergent adverse events. When a trial examines several outcomes, repeated statistical testing can increase the chance of finding at least one apparently notable result by chance. General statistical practice may address this through a prespecified hierarchy or a multiplicity adjustment, but the approach depends on the trial's aims and analysis plan. The registry record does not report a multiplicity strategy, endpoint hierarchy beyond identifying the primary outcome, or adjustment method.

The record does not report p-values or effect estimates for the primary comparison, and it does not report analyses for the listed secondary outcomes. It states that no analysis count is recorded and identifies seven secondary outcomes. Therefore, the available record does not permit assessment of which secondary endpoints were analyzed, whether any were statistically significant, or how selective reporting may have affected interpretation. The number of secondary outcomes listed should not be taken as evidence that every one produced a result.

Missing-data information is also limited. For overall survival, the endpoint description specifies censoring at last contact for participants alive. It does not report how many participants were censored, lost to follow-up, or missing other outcome information. Handling of missing values for tumor response, quality of life, or the other secondary outcomes is not reported in the registry record.

The primary outcome population consists of participants who received at least 1 dose, but an all-randomized analysis is not reported. The registry lists enrollment of 36 and primary outcome arm counts of 15 and 16, without an accounting of participant disposition. This leaves the extent and consequences of exclusions or missing outcome data uncertain. The trial's original sample-size justification, interim analyses, and detailed statistical plan are also not reported in the registry record.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hepatic encephalopathy (12.5%), dyspnoea (6.2%), respiratory failure (6.2%), general physical health deterioration (6.2%), influenza like illness (6.2%).

The registry's adverse-event results list serious adverse events affecting 0 of 15 participants in the TheraSphere group and 12 of 16 participants in the sorafenib group. It lists other adverse events affecting 6 of 15 and 16 of 16 participants, respectively. The registry also lists deaths of 15 and 16 participants in the two groups. These are participant counts in the posted adverse-event results; they do not by themselves establish that an event was caused by the assigned treatment.

The registry identifies the following serious-event terms among its most frequently reported terms: Hepatic encephalopathy, 12.5%; Dyspnoea, 6.2%; Respiratory Failure, 6.2%; General Physical Health Deterioration, 6.2%; Influenza like illness, 6.2%; and Diarrhoea, 6.2%. Those percentages are reproduced as reported. The record does not provide the event-level context needed to determine timing, recurrence, severity grading, relatedness to treatment, or the denominator used for each term.

General statistical context: adverse-event counts are descriptive and depend on how events were collected, coded, and attributed. Counts of participants affected are not the same as counts of individual events, since one participant may experience more than one event. Differences in reported counts should be interpreted alongside follow-up duration and exposure, which are not described in enough detail in the registry record to make a comprehensive comparative safety assessment.

12. What the Result Does, and Does Not, Mean

The posted results establish that the registry records a randomized comparison of TheraSphere and sorafenib and reports median overall survival for participants who received at least 1 dose. The reported medians are 14.5 months and 18.2 months, with full ranges of 2.3 to 22.0 months and 2.9 to 18.2 months. These values describe the groups as reported; they are not a complete account of survival at every point in follow-up.

They do not establish that one treatment improves survival relative to the other. The registry does not provide the primary comparative estimate, its confidence interval, or a statistical test. Nor does it provide enough information to evaluate the extent of censoring, the number at risk over time, the impact of participants not represented in the treated population, or the robustness of the analysis. The observed medians cannot answer those questions by themselves.

General statistical explanation: a failure to report a statistically significant comparison is not the same as demonstrating equivalence or no difference. To support equivalence, a trial would ordinarily need a prespecified margin and an analysis designed to assess whether the uncertainty interval falls within that margin. Such information is not reported in the registry record. Conversely, a numerical difference between medians does not establish a genuine treatment effect when its uncertainty has not been quantified.

The safety results also require restraint. The registry reports participant counts with serious and other adverse events, plus counts of deaths, but these counts alone do not establish treatment causality or permit a complete benefit-risk comparison. Information on event timing, treatment exposure, relatedness, and follow-up is not reported in enough detail to resolve those questions.

The record's terminated status is relevant context, but the reason for termination is not reported. It cannot be assumed that termination was due to a safety concern, an efficacy finding, recruitment, or another cause. The soundest interpretation is therefore limited: the registry provides arm-specific descriptive survival and adverse-event data, but not the analysis details necessary for a firm comparative conclusion.

13. Why This Trial Matters Statistically

This trial is statistically notable because it combines random allocation with an endpoint that requires attention to follow-up and censoring. Randomization can support a balanced comparison of assigned treatments, while overall survival incorporates the timing of deaths and the partial follow-up of participants still alive. Interpreting the endpoint well therefore requires more than comparing two medians: readers also need information about uncertainty, censoring, follow-up, and the analysis population.

The registry record illustrates the distinction between a descriptive summary and an inferential comparison. It reports median overall survival and full ranges for each arm, but does not report a hazard ratio, confidence interval, or p-value. The result can be described accurately at the level of the posted medians, yet the strength and precision of the comparative evidence cannot be judged from those summaries alone. This distinction is important in clinical research, where numerical contrasts can appear meaningful without enough information to assess uncertainty.

The treated-population definition is another important statistical detail. The primary outcome is described for participants who received at least 1 dose, while the trial is identified as randomized and the record gives an enrollment total of 36. The arm counts for the primary outcome are 15 and 16, but no participant flow explains the relationship among these figures. Transparent accounting matters because exclusions after randomization may affect the balance created by random allocation.

Finally, the record lists several secondary outcomes and marks the trial as terminated, but does not report a multiplicity strategy, detailed analysis plan, or reason for termination. These gaps do not invalidate the posted descriptive data, but they limit the conclusions that can be drawn about the overall trial evidence. The statistical lesson is to match the strength of interpretation to the detail actually reported.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several trial-specific gaps constrain interpretation. The registry reports enrollment of 36, but the primary outcome population includes 15 participants in the TheraSphere arm and 16 in the sorafenib arm. It does not provide a participant flow diagram or narrative explaining the difference, so the number and reasons for any exclusions from the survival population cannot be assessed.

The trial is described as randomized and open-label, but the registry record does not report the randomization ratio, sequence-generation method, or allocation concealment. It also does not report baseline characteristics by arm, which would help readers assess the groups' observed comparability. The primary outcome population is defined by receipt of at least 1 dose, and an intention-to-treat analysis is not reported.

The primary survival results lack a hazard ratio, confidence interval, p-value, survival curve, number-at-risk information, and details of the censoring pattern. The reported full ranges are not uncertainty intervals. Without these elements, neither the precision of the medians nor the comparative treatment effect can be evaluated adequately. The registry also does not describe the statistical model used to estimate survival summaries.

The study is marked terminated, but the reason is not reported. The registry does not describe an interim analysis, stopping rules, the trial's original sample-size justification, or whether the analysis plan changed. Although secondary outcomes are listed, corresponding analyses are not reported in the record. Multiplicity handling and missing-data methods for those outcomes are also not reported.

Safety interpretation is limited by the absence of detailed event timing, duration of treatment exposure, event-level counts, and attribution to treatment. These omissions mean that the adverse-event participant counts are useful as descriptive registry results but do not provide a complete basis for comparing safety between arms.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Efficacy Evaluation of TheraSphere to Treat Inoperable Liver Cancer With Blockage of the Portal Vein — NCT01887717.
• Registry record last updated 2021-04-21; status last verified 2021-04.
• Results first posted 2019-10-29.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The registry describes a prospective, randomized, open-label, parallel-group Phase 3 trial comparing TheraSphere with sorafenib for inoperable, advanced hepatocellular carcinoma with portal vein thrombosis. It records enrollment of 36, a start date of 2014-02-27, a primary completion date of 2017-05-23, and terminated status. The reason for termination is not reported in the registry record.

The primary endpoint is overall survival from randomization, with censoring at last contact for participants alive. The treated-population medians are 14.5 months for TheraSphere and 18.2 months for sorafenib. The record does not report a hazard ratio, confidence interval, p-value, or detailed survival analysis. It lists seven secondary outcomes but does not report their analyses.

The posted safety results include participant counts for serious and other adverse events and deaths. They do not establish event causality or provide a complete comparative safety assessment. Key details on participant disposition, missing data, multiplicity, the original sample-size justification, and the statistical analysis plan are not reported in the registry record.