Neo ALTTO (Neoadjuvant Lapatinib and/or Trastuzumab Treatment Optimisation) Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neo ALTTO (Neoadjuvant Lapatinib and/or Trastuzumab Treatment Optimisation) Study (NCT00553358), a randomized phase 3 study. The primary outcome is Number of Participants With Pathological Complete Response (pCR) at the Time of Surgery.

2. Clinical Question

The study asked whether pathological complete response at surgery differed among three neoadjuvant treatment strategies for women with HER2/ErbB2-positive primary breast cancer. Each strategy included targeted treatment before surgery and paclitaxel during the later neoadjuvant period. One used lapatinib, one trastuzumab, and one combined lapatinib with trastuzumab.

The central statistical contrast is whether the observed pCR counts differ between groups, and how precisely those differences are estimated. The randomized design supports comparisons between assigned groups, but the registry record alone does not establish how every aspect of trial conduct, outcome assessment, or analysis was handled. It also does not report the original statistical design assumptions or a complete account of multiplicity and missing-data procedures.

3. Trial Design

Neo ALTTO was a randomized, multicenter, open-label Phase III study sponsored by Novartis Pharmaceuticals. The registry lists enrollment of 455 participants and a completed study. Participants were assigned to one of three neoadjuvant strategies: lapatinib, trastuzumab, or lapatinib plus trastuzumab. Paclitaxel was added after the initial treatment period and continued until definitive surgery.

The registry describes a six-week biological window followed by twelve weeks of targeted therapy with weekly paclitaxel before surgery. In the combination group, the lapatinib and trastuzumab doses changed during the paclitaxel period. After surgery, the described treatment included three courses of adjuvant FEC followed by further targeted therapy. These post-surgical treatments form part of the treatment strategy, but the primary endpoint was the pathological response measured at surgery.

The study was open label, meaning treatment assignment was not masked according to the registry. The record identifies the primary analysis population as intent-to-treat: all randomized participants except those who withdrew consent for use of any data before receiving study medication. The primary endpoint timeframe was weeks 20 to 22. Its definition was based on examination of surgical breast and axillary node resection specimens, with the stated response criterion focused on invasive cancer in the breast specimen.

The record does not report the original sample-size justification, the planned statistical analysis details, interim analyses, or the procedures used to handle missing primary endpoint information. Those omissions limit what can be concluded about the trial’s original design choices. The planning illustrations discussed separately are based on the observed comparisons and are not a substitute for the trial’s own design assumptions.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-Treat (ITT) Population: all participants randomized to treatment, except for those who withdrew their consent to use any of their data (permitted by law in certain countries) prior to receiving any study medication

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 Number of Participants With Pathological Complete Response (pCR) at the Time of Surgery, measured over Weeks 20 to 22.

The primary endpoint was the number of participants with pathological complete response (pCR) at surgery, assessed during weeks 20 to 22. The registry defines pCR as no invasive cancer in the breast or only non-invasive in situ cancer in the breast specimen. Surgical breast and axillary node specimens were evaluated according to National Surgical Adjuvant Breast and Bowel Project guidelines. The description specifically states that those guidelines do not take histological nodal status into account.

Pathological response is used in this study because tissue removed during surgery can be examined for residual invasive cancer in the breast. It therefore provides a defined outcome at a specified treatment milestone, rather than relying only on a change in tumor size or a clinical examination. The registry’s primary measure is a count of participants meeting the stated response definition, so the group comparison concerns the proportion of participants with that outcome, even though the posted endpoint itself is expressed as a number.

The endpoint has boundaries that matter for interpretation. Its definition is based on the breast specimen and the stated pathology criteria; it should not be broadened into a claim about all disease sites or long-term outcomes. The record also lists event-free survival as a secondary outcome, but it does not make that the primary endpoint. A pCR comparison therefore answers a specific question about pathological findings at surgery, not every question about subsequent clinical course.

7. Results Posted in the Registry

• Trastuzumab 2 mg/kg: 29.5% (44/149) versus 24.7% (38/154) on control; risk difference 4.9 percentage points (95% CI -5.1 to 14.9), risk ratio 1.20 (0.83–1.73); unadjusted two-proportion z-test p = 0.342.
• Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg: 51.3% (78/152) versus 24.7% (38/154) on control; risk difference 26.6 percentage points (95% CI 16.2 to 37.1), risk ratio 2.08 (1.52–2.85); unadjusted two-proportion z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Number of Participants With Negative Lymph Nodes at the Time of Surgery

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Trastuzumab 2 mg/kg: 58.6% (82/140) versus 51.8% (72/139) on control; risk difference 6.8 percentage points (95% CI -4.9 to 18.4), risk ratio 1.13 (0.91–1.40); unadjusted two-proportion z-test p = 0.255.
  • Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg: 73.0% (100/137) versus 51.8% (72/139) on control; risk difference 21.2 percentage points (95% CI 10.0 to 32.3), risk ratio 1.41 (1.17–1.70); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Trastuzumab 2 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants Starting Paclitaxel Before Completing 6 Weeks of Treatment With Either Lapatinib or Trastuzumab

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Trastuzumab 2 mg/kg: 8.2% (12/146) versus 5.4% (8/149) on control; risk difference 2.9 percentage points (95% CI -2.9 to 8.6), risk ratio 1.53 (0.64–3.64); unadjusted two-proportion z-test p = 0.330.
  • Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg: 4.0% (6/149) versus 5.4% (8/149) on control; risk difference -1.3 percentage points (95% CI -6.1 to 3.5), risk ratio 0.75 (0.27–2.11); unadjusted two-proportion z-test p = 0.584.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Trastuzumab 2 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With the Indicated Biomarker Expression - PIK3CA.

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Trastuzumab 2 mg/kg: 18.8% (21/112) versus 23.4% (29/124) on control; risk difference -4.6 percentage points (95% CI -15.0 to 5.7), risk ratio 0.80 (0.49–1.32); unadjusted two-proportion z-test p = 0.384.
  • Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg: 25.2% (30/119) versus 23.4% (29/124) on control; risk difference 1.8 percentage points (95% CI -9.0 to 12.6), risk ratio 1.08 (0.69–1.68); unadjusted two-proportion z-test p = 0.740.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Trastuzumab 2 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lapatinib 1000/750 mg + Trastuzumab 2 mg/kg and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists a range of secondary outcomes, including pathological and surgical findings, tumor-size change, treatment timing, and event-free survival. The posted secondary comparisons available in the record show a pattern in which the combination group’s negative-node count is higher than the lapatinib group’s, while the trastuzumab-versus-lapatinib comparison is less conclusive. The paclitaxel-timing comparison does not show a clear difference between the groups in the unadjusted results. The baseline biomarker measure is descriptive of group composition, rather than a post-treatment outcome.

This broad pattern is directionally consistent with the primary pCR comparison of combination therapy against trastuzumab, but secondary outcomes do not replace the primary endpoint. Negative lymph-node status, treatment timing, and baseline biomarker expression address different questions from pCR in the breast specimen. Event-free survival is listed among secondary outcomes, but details needed to interpret its results are not included in the available endpoint summaries. The registry record therefore does not support a full synthesis of every secondary outcome.

Secondary results often carry less inferential weight because many outcomes and contrasts can be examined. If each is tested without accounting for the overall number of tests, the chance of at least one apparently unusual result can increase. The registry record does not report a multiplicity adjustment for secondary endpoints or a hierarchy that would distinguish confirmatory from exploratory analyses. The listed secondary findings should consequently be treated as supportive or hypothesis-generating, not as independent confirmation of the primary finding.

Some secondary outcomes also have different denominators from the primary analysis, reflecting the participants represented in those specific measures. Without a complete account of missingness and analysis procedures, differences in denominators should be noted rather than explained by assumption. The secondary outcomes can add context, but they do not establish effects on outcomes that are not reported.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Participants With Overall Response at Week 6
• Overall Response at the Time of Surgery
• Number of Participants With Negative Lymph Nodes at the Time of Surgery
• Number of Participants With Actual Indicated Surgery
• Mean Change From Baseline in Tumor Size at Week 6 and at Surgery
• Number of Participants Starting Paclitaxel Before Completing 6 Weeks of Treatment With Either Lapatinib or Trastuzumab
• Event-free Survival (EFS) - Median Clinical Follow-up
• Event-free Survival (EFS) - Events and Censoring

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary for each participant: the pathology assessment either meets the registry’s pCR definition or it does not. A binomial analysis is a natural way to describe a binary outcome because it models the number of participants with the event relative to the number assessed. The record’s posted group results are counts, and the comparisons can be expressed as differences in event proportions. These are distinct from comparisons of mean values, which would be relevant for a continuous outcome.

What does a risk difference describe?

A risk difference compares the event proportion in one group with the event proportion in another. A positive value means the event proportion is higher in the group listed first in that contrast; a negative value means it is lower. The value is expressed on the proportion scale, often presented as a percentage-point difference. It is an absolute comparison, unlike a risk ratio, which expresses one group’s event proportion relative to another. The registry’s recalculated comparisons use risk differences, while the sponsor’s posted analyses report percentage-point differences.

What does a confidence interval add?

A confidence interval communicates uncertainty around an estimate. It reflects both the observed contrast and the amount of information available for that contrast. A wider interval means a less precise estimate; a narrower interval means greater precision under the analysis. If an interval includes no difference, values consistent with no group difference remain among the plausible values covered by that interval. If the interval lies entirely on one side of no difference, the estimate is more directionally consistent. The interval is not a guarantee that the true value lies within its bounds, and it does not address bias or missing information that the analysis does not account for.

How should p-values be read?

A p-value describes how compatible the observed data are with a specified no-difference model, under the assumptions of the test. It is not the probability that the treatment effect is absent, nor does it measure the size or importance of an effect. A small p-value can accompany an imprecise or modest estimate, and a larger p-value does not prove the groups are equivalent. The registry includes sponsor p-values for two primary pairwise comparisons and also lists unadjusted recalculated p-values. Because the reported and recalculated analyses differ, the sponsor’s own analysis takes precedence when describing the trial’s reported findings.

Why distinguish sponsor analyses from recalculations?

The recalculations are simple, unadjusted analyses from the posted group summary counts. They do not reproduce any unreported adjustment, stratification, or other choices the sponsor may have used. The registry record flags a discrepancy between the recalculations and sponsor results. Consequently, recalculated values are useful as a transparent secondary check on the posted summaries, but they should not replace the trial’s own reported analysis. The available record does not explain the source of the discrepancy, so attributing it to a particular method would be speculation.

What does unadjusted mean?

An unadjusted comparison is calculated without accounting for other variables or applying a stated correction for the family of comparisons. This is not necessarily inappropriate for a simple descriptive contrast, but it can make results harder to interpret when many outcomes or pairwise tests are examined. The registry record does not report a multiplicity procedure for the secondary endpoints or a complete statistical analysis plan. Therefore, secondary unadjusted p-values should be treated as exploratory signals rather than definitive evidence.

Why do sample-size illustrations use different effect sizes?

A planning calculation asks how many participants would be needed under stated assumptions to detect a target effect with a specified probability, called power. The posted illustrations use the observed effect and fractions of it, alongside a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. For a smaller target effect, the groups generally need more information to distinguish that effect from random variation. These calculations are hypothetical planning illustrations based on the observed contrasts; they are not the trial’s original design calculation.

Statistical significance and precision are related but not interchangeable. A p-value summarizes a test against a particular null hypothesis, while an interval makes the range of estimates compatible with the data easier to see. Both depend on the analysis assumptions. Neither accounts automatically for unreported design features, outcome ascertainment limitations, missing-data mechanisms, or multiplicity. Those aspects require methods information that is not reported in the registry record.

11. Confidence Intervals and Precision

The sponsor’s primary comparisons have notably different levels of precision. For lapatinib compared with trastuzumab, the sponsor reported a difference of -4.85 percentage points and a 97.5% confidence interval from -17.6 to 8.16. Because this interval crosses no difference, the data are compatible with estimates in either direction. The result is not precise enough to establish a clear direction for this pairwise contrast.

For trastuzumab compared with the combination, the sponsor reported a difference of 21.79 percentage points and a 97.5% confidence interval from 9.08 to 34.23. The interval excludes no difference and supports a higher estimated pCR frequency in the combination group in this analysis. Its width still indicates uncertainty about the size of the difference: the endpoints represent meaningfully different possible magnitudes of the contrast.

Confidence intervals are more informative than a binary significant/not-significant label because they show both direction and a range of estimates. They do not tell whether an effect is important to patients, whether it persists, or whether the analysis is free of bias. The registry record does not report a threshold for a clinically meaningful pCR difference, so the intervals cannot be classified against such a threshold.

Precision also depends on the number of participants and the number of observed outcomes. The primary groups include 154.0, 149.0, and 152.0 participants, and the counts with pCR are 38.0, 44.0, and 78.0. These group sizes provide the context for reading the intervals, but they do not eliminate uncertainty. The sponsor’s 97.5% intervals are the intervals reported for its analyses; the reason for that interval level is not reported in the registry record.

The unadjusted recalculations have their own intervals and are not interchangeable with the sponsor’s. Given the recorded discrepancy, the sponsor’s intervals take precedence as the trial’s reported inferential results. In either case, precision should be judged from the interval rather than inferred from the number of decimal places in an estimate.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size illustrations translate each observed comparison into hypothetical future planning requirements. They use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the observed effect or at a stated fraction of it. The registry record does not report the trial’s original sample-size assumptions, so these illustrations should not be read as an account of how the study was planned.

For the binary outcomes presented, the figures are participants per arm, not totals across both groups. The columns for 80% power and 90% power describe the number of participants per arm needed, under the stated assumptions, for an analysis to have the corresponding probability of detecting the target effect if that effect were present. The higher power target generally requires a larger planned group size. The rows for 100%, 75%, and 50% of the observed effect show how the planning requirement changes when the target difference is made smaller.

Smaller effects are harder to distinguish from random variation, so they generally require more participants. The requirements can therefore rise substantially when planning for a fraction of the observed effect, particularly when the observed comparison is small. These are planning illustrations, not observed power, and they do not show the probability that the trial’s findings are true. They also do not account for design features beyond the stated assumptions.

The sample-size curve should be read in the same way: it represents hypothetical planning needs as the target effect changes, rather than additional observed trial results. Its values depend on the assumptions noted above and on the observed effect used as the starting point. They should not be taken as a recommendation for a future study without a separate design assessment.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.247; treatment proportion p₁ = 0.295; average p̄ = 0.271.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6286 = 1.2320.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6277 = 0.5283.
4. n = (1.2320 + 0.5283)² / (0.0485)² = 1,314.7, rounded up to 1,315 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.246753246753247                          # control proportion
p1 <- 0.295302013422819                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(553358)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 1,315 (80% power), 1,760 (90% power)
 75% of observed effect: n per arm = 2,305 (80% power), 3,085 (90% power)
 50% of observed effect: n per arm = 5,111 (80% power), 6,841 (90% power)
power.prop.test: 1314.7 
Simulated power at n = 1,315 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry reports sponsor analyses for two primary comparisons: lapatinib versus trastuzumab, and trastuzumab versus the combination. For the first, the sponsor reported a -4.85 percentage-point difference, a 97.5% confidence interval from -17.6 to 8.16, and a p-value that was not small. For the second, the sponsor reported a 21.79 percentage-point difference, a 97.5% confidence interval from 9.08 to 34.23, and p < 0.001.

Simple unadjusted recalculations from the posted summary counts do not exactly match the sponsor’s results; the registry record flags this discrepancy. The reason for the difference is not reported in the registry record. It would not be appropriate to guess whether it reflects a particular adjustment, analysis population, or other method. The sponsor’s own analysis takes precedence when describing the trial’s reported inferential findings.

The recalculation remains a useful check on the direction and broad pattern of the posted counts, but it is not a substitute for the sponsor analysis. The sponsor’s analysis for lapatinib versus the combination is not reported in the registry record. Comparisons beyond the sponsor’s two reported contrasts should therefore be identified as recalculated or descriptive rather than attributed to the sponsor.

15. Multiplicity, Missing Data and Analysis Populations

The registry record does not report whether the primary comparisons were adjusted for multiplicity, how the two sponsor-reported comparisons were prioritized, or whether a formal testing hierarchy was prespecified. It also does not report a multiplicity procedure for secondary outcomes. Since the study includes multiple treatment groups and a broad set of secondary measures, interpreting isolated secondary p-values without this information requires caution. The stated 97.5% confidence intervals for sponsor analyses are reported as such; the record does not explain the rationale for that interval level.

The primary population is described as intent-to-treat, including all randomized participants except those who withdrew permission to use any data before receiving study medication. This definition is important because it identifies the intended population for the pCR analysis, but it does not describe how participants with unavailable pathology results were handled. The missing-data method, the number of participants missing the primary outcome, and any sensitivity analyses are not reported in the registry record.

The record also does not report the original statistical analysis plan, the trial’s own sample-size assumptions, interim analyses, or whether any analyses were adjusted for baseline characteristics or other design features. These details could affect interpretation of the sponsor’s inferential results, but their absence does not justify guessing which methods were used. The discrepancy between sponsor results and the simple recalculations is flagged, while its cause is not explained.

Open-label assignment can be relevant to how outcomes are assessed and how treatment is delivered. The primary endpoint uses surgical pathology, which is a defined assessment, but the registry record does not state whether pathologists were masked to treatment assignment or describe quality-control procedures. That information is not reported in the registry record. These uncertainties are reasons to keep conclusions close to the stated endpoint and reported analysis rather than extending them beyond the documented evidence.

16. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hypertransaminasaemia (15.2%), neutropenia (10.8%), diarrhoea (6.0%), febrile neutropenia (5.4%), pyrexia (3.4%).

The registry lists serious adverse events and other adverse events by treatment group, with different numbers of participants at risk in the safety summaries. In the combination group, 61 of 149 participants were reported with a serious adverse event, and 147 of 149 with other adverse events. In the lapatinib group, the corresponding counts were 58 of 151 and 148 of 151. In the trastuzumab group, they were 36 of 148 and 141 of 148. The listed deaths were 1, 2, and 0, respectively.

The most frequent serious adverse event terms listed include hypertransaminasaemia at 15.2%, neutropenia at 10.8%, diarrhoea at 6.0%, febrile neutropenia at 5.4%, pyrexia at 3.4%, and hyperbilirubinaemia at 2.7%. The record does not give an arm-specific breakdown for these listed terms in the available safety summary. These figures describe reported events; they do not by themselves establish that a treatment caused an event or permit a complete comparison of risk.

Safety counts should be read with their denominators and definitions. Serious and other adverse events are separate registry categories, but the record does not explain overlap between categories or provide enough detail to characterize duration or severity beyond the serious-event designation. The counts are descriptive and do not replace a full safety analysis.

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

The primary results describe pathological findings in breast specimens at surgery for participants assigned to three neoadjuvant strategies. The sponsor-reported comparison between trastuzumab and the combination estimates a higher pCR frequency in the combination group, with a confidence interval that excludes no difference. The comparison between lapatinib and trastuzumab is less conclusive because its reported interval includes no difference. These are findings about the specified endpoint and contrasts.

They do not establish that the combination is superior for every outcome, every participant, or every treatment setting. The registry record does not provide a sponsor-reported direct comparison between lapatinib and the combination. Nor should the pCR result be extended automatically to event-free survival, long-term outcomes, quality of life, or other endpoints. Event-free survival appears among the secondary outcomes, but its detailed results are not available in the record.

A confidence interval describes statistical uncertainty under an analysis, not the full uncertainty about bias, generalizability, or study conduct. The record identifies an open-label randomized design and an intent-to-treat population definition, but it does not report all details of outcome assessment, missing data, multiplicity, or original statistical planning. The discrepancy between simple recalculations and sponsor analyses is an additional reason to distinguish trial-reported results from unadjusted checks.

The results also do not provide medical advice. They are evidence about pCR counts and the sponsor’s stated comparisons in this particular study population. Whether a difference is clinically meaningful cannot be determined from statistical significance alone, and the registry record does not report a threshold for meaningful pCR change. A sound reading keeps the magnitude, interval, outcome definition, and limitations in view without turning one endpoint into a broader claim about treatment benefit.

18. Why This Trial Matters Statistically

Neo ALTTO illustrates why reporting an effect estimate and confidence interval alongside counts is important. The primary outcome is a count of participants meeting a pathology-based definition, but the scientific comparison concerns how often that outcome occurred across randomized groups. The contrast between the two sponsor-reported comparisons shows that one result can be compatible with a range spanning no difference while another has an interval on one side of no difference. A single label such as significant or not significant would conceal that distinction in precision and direction.

The study also highlights the importance of identifying whose analysis is being described. Simple calculations from group summaries can be transparent and useful, but they may not reproduce the sponsor’s analysis. When a discrepancy is flagged, the registry’s own analysis should take precedence as the reported trial result, while the recalculation remains clearly identified as a secondary, unadjusted check. This distinction prevents a derived comparison from being mistaken for the trial’s prespecified analysis.

Finally, the study’s multiple outcomes make the role of multiplicity and analysis plans central to interpretation. A collection of secondary findings can provide context, but without a reported testing strategy, the strength of those findings is difficult to calibrate. The case demonstrates why registry entries should be read not only for results, but also for what they do not report: design assumptions, missing-data handling, and the plan for interpreting multiple comparisons.

19. 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 limitations are specific to the available record for this trial. The study was open label, and the registry does not state whether the pathology assessors were masked to assigned treatment. Because pCR is based on surgical specimens, the assessment method is defined, but the record does not describe assessment blinding or quality-control procedures.

The primary endpoint’s stated definition focuses on invasive disease in the breast specimen and notes that the cited pathology guidelines do not account for histological nodal status. Therefore, the pCR label should be interpreted according to that precise definition, rather than assumed to mean absence of invasive disease at every site. Other surgical outcomes, including negative lymph nodes, are listed separately.

The registry reports enrollment of 455, while the primary analysis groups contain 154.0, 149.0, and 152.0 participants. It does not report a complete explanation for the difference between enrollment and the group denominators used for the primary endpoint. The analysis population definition excludes participants who withdrew consent to use their data before receiving any study medication, but missing endpoint handling is not reported.

The record flags a discrepancy between sponsor analyses and simple recalculations without explaining its source. It also does not report the original sample-size justification, a full statistical analysis plan, the handling of multiplicity, or interim-analysis procedures. These omissions constrain the interpretation of secondary findings and the ability to reproduce the sponsor’s methods. They do not erase the posted results, but they limit conclusions beyond the comparisons and endpoint explicitly reported.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Neo ALTTO (Neoadjuvant Lapatinib and/or Trastuzumab Treatment Optimisation) Study — NCT00553358.
• Registry record last updated 2021-09-21; status last verified 2021-08.
• Results first posted 2011-10-13.
• 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.

23. Record Summary

Neo ALTTO was a completed randomized, multicenter, open-label Phase III study sponsored by Novartis Pharmaceuticals, with 455 participants enrolled. It compared neoadjuvant lapatinib, trastuzumab, and their combination with paclitaxel in women with HER2/ErbB2-positive primary breast cancer. The primary endpoint was the number of participants with pCR at surgery during weeks 20 to 22.

The sponsor reported a less conclusive pCR comparison between lapatinib and trastuzumab, and a higher estimated pCR frequency for the combination compared with trastuzumab. The registry flags a discrepancy between sponsor results and simple unadjusted recalculations; the sponsor’s own analysis takes precedence. Important details, including the original sample-size justification, missing-data procedures, and multiplicity plan, are not reported in the registry record.