Sorafenib Induced Autophagy Using Hydroxychloroquine in Hepatocellular Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Sorafenib Induced Autophagy Using Hydroxychloroquine in Hepatocellular Cancer (NCT03037437), a non-randomized phase 2 study. The primary outcome is Time to Tumor Progression Evaluated Via Tumor Imaging.

2. Clinical Question

3. Trial Design

This is a phase 2 study listed as completed, with non-randomized allocation, no masking, and a parallel model. The registry lists enrollment of 37. Its stated purpose is to study whether SOR with HCQ has improved efficacy compared with SOR alone and whether adding HCQ after progression on SOR could lead to disease stability. Those aims do not make the treatment groups randomized or directly comparable: the registry explicitly cautions that the arms are not a like-for-like control comparison.

The descriptions distinguish participants with no prior systemic treatment from those progressing on SOR. In the first group, SOR begins at 400 mg by mouth twice daily on C1D1. The clinician is to reduce the SOR dose based on toxicity on C1D15, when HCQ 400 mg by mouth daily is to start. The registry’s wording ends this description with “C2D1”; it does not clarify further what happens at that point. In the progression group, HCQ is added to the SOR dose tolerated at progression.

The registry lists a start date of 2017-02-16 and primary completion of 2025-06-27. It reports overall status as completed. The primary endpoint timeframe is from baseline to study completion, approximately 1 year. These dates and status describe the study record; they do not supply a primary outcome estimate. The registry record does not report further details of the allocation process, treatment adherence, changes in exposure, or a statistical analysis plan.

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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Participants were considered to be evaluable if they completed cycle 1 (approximately 4 weeks). No participants were evaluable for Cohort 2, since there were no participants that completed cycle 1 (4 weeks) in this cohort.

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.

6. Primary Endpoint

The registry lists the primary outcome as Time to Tumor Progression Evaluated Via Tumor Imaging, measured over From baseline to study completion, approximately 1 year.

The primary endpoint is time to tumor progression evaluated through tumor imaging, with the result parameter described as a mean and dispersion as a 95% confidence interval. The timeframe is from baseline to study completion, approximately 1 year. The registry says that CAT scans or MRI are performed at screening and then every other cycle, with tumor evaluation using RECIST version 1.1.

Imaging-based assessment provides a defined way to track tumor progression over repeated evaluations. In general, a time-to-progression endpoint represents the interval from a specified starting point to the recorded progression event; the registry identifies baseline as the starting reference. The posted endpoint description does not specify how progression dates were assigned between scans, how deaths or missed scans were handled, or how censoring was defined. Those details are not reported in the registry record.

The record states that participants were evaluable if they completed cycle 1, approximately 4 weeks. No participants in the progression group completed cycle 1, so none were evaluable for that cohort. The registry record does not report an observed mean time to progression or its confidence interval. The endpoint definition therefore describes what was intended to be measured, rather than the size or precision of an observed treatment effect.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is described as a continuous measure summarized by a mean and a 95% confidence interval. A mean gives a central summary of measured values; a confidence interval conveys statistical uncertainty around an estimate under the model and assumptions used. In general, a wider interval indicates less precision than a narrower interval, but the interval’s actual width and the underlying mean are not reported in the registry record. No conclusion about the degree of precision can therefore be drawn for this trial’s primary endpoint.

What does the imaging schedule contribute?

The registry describes CAT scans or MRI at screening and then every other cycle, with response evaluation according to RECIST version 1.1. Repeated imaging offers planned opportunities to identify tumor progression during follow-up. In general, an event detected at a scan occurred by that assessment, but the exact time of progression may fall between imaging visits. The registry does not report how an event date was assigned in that circumstance, whether imaging schedules varied, or how missed or delayed assessments were handled. Those choices can affect the measured time to progression and its analysis.

What does the evaluability rule mean?

The record defines evaluable participants as those who completed cycle 1, approximately 4 weeks. This is a post-enrollment condition: being enrolled does not by itself establish that a participant contributed to the primary endpoint population. The record explicitly says no participants in the progression group completed cycle 1, so none in that cohort were evaluable. It does not report how many participants in the other group completed the cycle or how participants who did not complete it were treated in any analysis. A complete description of the analysis population and missing outcome handling is therefore unavailable.

How should the planning illustrations be read?

The registry’s sample-size assumptions describe planning illustrations computed from the observed effect, not the trial’s own design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is expressed as a stated fraction of the observed effect. Such illustrations can show, in general, how hypothetical sample sizes relate to an assumed effect and statistical criteria. They are not evidence that the study enrolled the number required for a prespecified target, and they are not observed power. The registry record does not report the original sample-size justification, the planned target enrollment, or a protocol-based power calculation.

The record describes the primary measure as a mean, but it does not provide details of the statistical model used to estimate it. It does not say whether the reported mean was intended to be calculated only among evaluable participants, whether adjustments for baseline characteristics were planned, or how censoring and incomplete follow-up would enter the analysis. These are important distinctions: a simple mean of observed times and an analysis that accounts for censored follow-up answer different statistical questions. The registry record does not report which approach was used.

In general, confidence intervals and significance tests depend on assumptions about the data and analysis. A normal approximation, as mentioned for planning illustrations, is an assumption for those illustrations and should not be presumed to describe the primary analysis. The registry record does not report an analysis method, test statistic, p-value, or confidence interval for the primary result. It also reports no analysis entries. Consequently, the statistical methods for the primary endpoint cannot be reconstructed from the posted record, and planning assumptions should not be confused with the analysis ultimately performed.

The study is non-randomized and has no masking. In general, random assignment can help balance measured and unmeasured characteristics across groups, while masking can limit some assessment or treatment-related influences. Neither feature is present as described in this record. That does not by itself determine the direction or size of bias, but it limits the strength of causal comparisons that can be made from group contrasts. The different treatment histories of the listed cohorts add a further interpretive concern: a contrast between them would not isolate the effect of adding HCQ alone.

The primary endpoint is clinically structured around imaging-defined progression, but the measurement plan alone is not a result. The record gives no numerical endpoint estimate, no confidence interval, and no account of the analysis dataset. The appropriate statistical reading is therefore descriptive: the registry specifies an imaging schedule, an endpoint summary, and an evaluability condition, while leaving the performed primary analysis and its results unreported.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and zero analysis entries. It does not report a plan for multiple comparisons, a hierarchy of endpoints, or adjustment for testing more than one hypothesis. Without a reported analysis plan, it is not possible to determine whether multiplicity was considered in the trial’s statistical interpretation. The absence of such details in the record is not evidence that no analyses were performed; it means their methods and results are not reported in the registry record.

The record gives an evaluability condition for the primary outcome: participants had to complete cycle 1, approximately 4 weeks. It also states that no participants in the progression group completed cycle 1, so none were evaluable in that cohort. The record does not report the evaluable number for the other group or explain whether participants who did not complete cycle 1 were excluded, included in another analysis population, or handled in another way.

Missing-data handling is not described. The registry does not state how incomplete imaging, missed visits, withdrawal, death, or other reasons for unavailable outcome data were handled. Nor does it report a sensitivity analysis examining the effect of missing observations. In general, missing outcomes can affect both an estimate and its uncertainty, particularly if the reasons for missingness relate to a participant’s outcome or treatment course. The posted record does not provide enough information to assess that possibility in this study.

Because the progression cohort has no evaluable participants under the stated rule, the primary endpoint cannot be interpreted for that cohort using the described evaluable population. That is a limitation of the available endpoint information, not a numerical estimate of the cohort’s outcome.

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: myocardial infarction (3.1%).

The registry reports serious adverse event figures for the group with no prior systemic treatment and for Cohort 2, identified as addition of HCQ. In the first group, 1 of 32 participants at risk had a serious adverse event; in Cohort 2, 0 of 5 had a serious adverse event. The record lists myocardial infarction as a top serious term at 3.1%. It also reports other adverse events in 28 of 32 participants in the first group and 2 of 5 in Cohort 2, and reports 1 death in each group.

These counts are descriptive. The registry record does not report whether events were judged related to SOR or HCQ, their timing, severity beyond the serious-event classification, or the definitions used for each category. The counts should not be treated as a comparison of treatment risks: the groups are non-randomized, differ by treatment history, and are not a like-for-like control comparison. The record does not explain whether event categories overlap, so the figures should not be added together. The listed death counts do not establish cause or treatment attribution.

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

The registry record documents a phase 2 study of SOR and HCQ in hepatocellular cancer, with a primary endpoint focused on imaging-assessed time to tumor progression. It describes a group with no prior systemic treatment and a group progressing on SOR, including how HCQ was added in each treatment context. It also records enrollment of 37 and the evaluability requirement for the primary outcome.

What the posted information does not provide is the primary outcome estimate. There is no reported mean time to progression, 95% confidence interval, or comparative primary analysis. Without these, the treatment effect’s size and precision cannot be described. In particular, the absence of evaluable participants in the progression group does not indicate that HCQ prevented progression or that progression did not occur; it means no participants in that cohort completed the stated minimum cycle for evaluability.

The design also limits causal interpretation. Participants were not randomly allocated, and the groups represent different treatment histories. The registry cautions that they are not a like-for-like control comparison. Even if numerical outcomes were available, a difference between such groups would not by itself separate treatment effects from differences associated with prior therapy, disease course, or other characteristics. The record does not provide adjusted comparative estimates or enough detail to evaluate confounding.

The safety figures report event counts and a serious-event term, but they do not establish whether an event was caused by either drug. The record does not provide the event attribution, timing, or a controlled safety comparison. The most supportable conclusion is therefore about what the registry records: the study’s intended question and endpoint, its non-randomized structure, the stated evaluability limitation, and descriptive safety counts. It does not support a quantified claim about efficacy or a causal claim about safety.

13. Why This Trial Matters Statistically

This trial illustrates why a registered research question and an interpretable treatment effect are not the same thing. The record describes a comparison of SOR with SOR plus HCQ as an aim, but the actual groups are non-randomized and organized around treatment history. The record explicitly warns that the groups are not like-for-like controls. That distinction is central to statistical interpretation: a group contrast in this setting may combine treatment differences with differences in who entered each group and why.

The endpoint also shows how analysis populations shape what can be learned. Completion of cycle 1, approximately 4 weeks, was required for evaluability, and no participants in the progression cohort completed that cycle. Thus enrollment alone does not tell how much information was available for the primary endpoint. The absence of an evaluable cohort limits the endpoint information, and should not be mistaken for a treatment outcome.

Finally, the registry record separates endpoint planning from reported evidence. It gives an imaging schedule, a mean and confidence-interval summary, and assumptions for planning illustrations, but it does not report the primary outcome estimate or an analysis. Those omissions prevent assessment of effect magnitude and precision. For clinical researchers and statisticians, the case underscores the importance of clearly reporting who contributed endpoint data, how incomplete follow-up was handled, what analysis was performed, and how uncertainty was quantified. Without those elements, the registry can describe study intent and conduct but cannot answer the primary statistical question.

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.

The main limitation for interpreting the primary endpoint is that its numerical result is not reported in the registry record. Although the measure is described as mean time to tumor progression with a 95% confidence interval, neither the mean nor the interval appears in the posted information. The record also lists zero analysis entries. It is therefore not possible to assess the effect estimate, its precision, or the analysis used to produce it.

The evaluability requirement narrows the population contributing to the primary endpoint. Participants had to complete cycle 1, approximately 4 weeks, and no participants in the progression cohort met that condition. The record does not report the evaluable count in the other group or provide participant flow that would explain completion, withdrawal, or other missing observations. It does not describe how missing scans or follow-up were handled.

Non-randomized allocation and the different treatment histories of the groups limit direct comparisons. The registry itself says the arms are not a like-for-like control comparison. The record does not report adjustment for potential confounding, details of allocation, or a comparative statistical model. No masking is reported, and the record does not explain any safeguards for imaging assessment.

Other details that would help evaluate the findings are also absent, including a full statistical analysis plan, the original sample-size justification, the definition and handling of progression dates between imaging visits, and the treatment attribution of adverse events. The registry’s planning assumptions are illustrations, not a substitute for those details or for primary outcome results.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Sorafenib Induced Autophagy Using Hydroxychloroquine in Hepatocellular Cancer — NCT03037437.
• Registry record last updated 2026-10-21; status last verified 2026-06.
• Results first posted 2026-10-21.
• 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

This completed phase 2 study, NCT03037437, evaluated sorafenib and hydroxychloroquine in hepatocellular cancer. It was non-randomized, unmasked, and parallel, with enrollment of 37. The groups were participants with no prior systemic treatment and participants progressing on sorafenib; the registry cautions that they are not a like-for-like control comparison.

The primary endpoint was mean time to tumor progression evaluated by CAT scan or MRI using RECIST version 1.1. Participants had to complete cycle 1, approximately 4 weeks, to be evaluable, and no participants in the progression cohort met that criterion. The registry record does not report the primary endpoint estimate or confidence interval. It provides descriptive serious and other adverse event counts, but does not report event attribution. The primary outcome’s magnitude and precision cannot be assessed from the posted record.